CN115099326A - Behavior prediction method, device, equipment and storage medium based on artificial intelligence - Google Patents

Abstract

The embodiment of the application belongs to the field of artificial intelligence, and relates to a behavior prediction method based on artificial intelligence, which comprises the following steps: acquiring user behavior data of a target user within a preset time period; performing characteristic construction on the user behavior data to generate a refueling behavior characteristic; the method comprises the steps of screening the refueling behavior characteristics based on a principal component analysis algorithm to obtain first characteristics, and screening the refueling behavior characteristics based on a preset characteristic selection algorithm to obtain second characteristics; determining a target fueling behavior characteristic based on the first characteristic and the second characteristic; and predicting the target refueling characteristic through the prediction model to generate a refueling behavior prediction result corresponding to the target user. The application also provides a behavior prediction device based on artificial intelligence, computer equipment and a storage medium. In addition, the application also relates to a block chain technology, and the refueling behavior prediction result can be stored in the block chain. The method and the device improve the processing efficiency and accuracy of the prediction of the refueling behavior intention of the user.

Description

Behavior prediction method, device, equipment and storage medium based on artificial intelligence

technical field

The present application relates to the technical field of artificial intelligence, and in particular, to a method, apparatus, device and storage medium for behavior prediction based on artificial intelligence.

Background technique

With the rapid development of Internet technology, more and more user behavior data has been recorded, and a large amount of online and offline data has brought possibilities for numerous applications. In response to the rigid demand for refueling services of car owners and users, some smart car applications have appeared. By introducing a new refueling mode of prepaid and non-inductive refueling in the application, car owners can enjoy their car life and save money for car owners. Considerable travel costs. In order to create a better refueling service, the existing marketing salesmen will conduct in-depth manual analysis of the historical user behavior data of the refueling service in the automotive application, and manually predict the car owner and user according to the service preference or service willingness of the car owner and user. Refueling intention in the future, and by issuing coupons to different types of car owners and users, guide car owners and users to enter the platform to enjoy services, so as to improve user loyalty and service experience to achieve efficient marketing. However, since the user's historical behavior data will become more and more accumulated over time, the existing processing method of manually predicting the future refueling intention of the car owner and user requires a lot of manpower and material resources, the processing efficiency is low, and there is no guarantee The accuracy of the forecast results.

SUMMARY OF THE INVENTION

The purpose of the embodiments of the present application is to propose a behavior prediction method, device, computer equipment and storage medium based on artificial intelligence, so as to solve the problem that the existing processing method of manually predicting the future refueling intention of vehicle owners and users requires a lot of manpower and material resources, The processing efficiency is low, and the accuracy of the prediction results cannot be guaranteed.

In order to solve the above-mentioned technical problems, the embodiment of the present application provides a behavior prediction method based on artificial intelligence, and adopts the following technical solutions:

Obtain user behavior data of target users within a preset time period;

Perform feature construction processing on the user behavior data, and generate fueling behavior features based on the user behavior data; wherein, the number of the fueling behavior features includes multiple;

The first feature is obtained by screening the refueling behavior feature based on a principal component analysis algorithm, and the second feature is obtained by screening the refueling behavior feature based on a preset feature selection algorithm;

determining a target fueling behavior feature corresponding to the user behavior data based on the first feature and the second feature;

The target refueling behavior feature is input into a preset prediction model, and prediction processing is performed on the target refueling feature through the prediction model to generate a refueling behavior prediction result corresponding to the target user.

Further, the step of screening the refueling behavior feature based on the principal component analysis algorithm to obtain the first feature specifically includes:

constructing a correlation coefficient matrix based on the refueling behavior feature;

Perform dimensionality reduction processing on the correlation coefficient matrix based on the principal component analysis algorithm to obtain a plurality of principal components;

Calculate the variance contribution rate of each of the principal components, sort all the variance contribution rates in descending order, and accumulate them in turn until exceeding the preset cumulative variance contribution rate threshold;

A designated principal component corresponding to the accumulation process is selected from all the variance contribution rates, and the feature corresponding to the designated principal component is used as the first feature.

Further, the step of determining the target fueling behavior feature corresponding to the user behavior data based on the first feature and the second feature specifically includes:

Merging the first feature and the second feature to obtain a corresponding feature set;

Perform matching processing on all the features in the feature set, and obtain a third feature that appears repeatedly in the feature set;

The third feature is used as the target fueling behavior feature.

Further, the prediction model includes an input layer, an LSTM layer, a fully connected layer and an output layer, and the target fueling behavior feature is input into a preset prediction model, and the target fueling feature is analyzed by the prediction model. The steps of performing prediction processing to generate a prediction result of refueling behavior corresponding to the target user specifically include:

Inputting the target refueling behavior feature into an input layer of the prediction model, and generating an output matrix corresponding to the target refueling behavior feature through the input layer;

Inputting the output matrix into the LSTM layer of the prediction model, and generating an output vector corresponding to the output matrix through the LSTM layer;

inputting the output vector into a fully connected layer of the prediction model, and generating a prediction classification result corresponding to the output vector based on a preset activation function in the fully connected layer;

The prediction classification result is input into the output layer of the prediction model, the prediction classification result is output through the output layer, and the prediction classification result is used as the refueling behavior prediction result corresponding to the target user.

Further, in the step of inputting the target refueling behavior feature into a preset prediction model, performing prediction processing on the target refueling feature through the prediction model, and generating a refueling behavior prediction result corresponding to the target user. Before, also included:

Obtain pre-collected fueling sample data;

Divide the refueling sample data into a training set and a test set according to a preset ratio;

Get a pre-built initial prediction model;

Based on the stochastic gradient descent algorithm, use the training set to train the initial prediction model to obtain a trained initial prediction model;

Validating the trained initial prediction model using the test set;

If the trained initial prediction model passes the verification, the trained initial prediction model is used as the prediction model.

Further, the step of using the test set to verify the trained initial prediction model and judging whether the verification passes, specifically includes:

obtaining the test set; wherein, the test set includes a plurality of test data, and classification information corresponding to each of the test data respectively;

Inputting the test data into the initial prediction model after training, and obtaining the first prediction classification result output by the initial prediction model after training and corresponding to each of the test data respectively;

determining, based on the classification information, a second prediction classification result that is correctly predicted among all the first prediction classification results;

calculating the quotient of the second predicted classification result and the first predicted classification result;

judging whether the quotient is greater than a preset threshold;

If it is greater than the preset threshold, it is determined that the trained initial prediction model has passed the verification; otherwise, it is determined that the trained initial prediction model has not passed the verification.

Further, the step of acquiring the user behavior data of the target user within a preset time period specifically includes:

Acquiring page user behavior data within the preset time period, and acquiring user information of the target user;

Based on the user information, filter out the user behavior data of the specified page corresponding to the user information from the page user behavior data;

The user behavior data of the specified page is preprocessed to obtain the user behavior data.

In order to solve the above-mentioned technical problems, the embodiment of the present application also provides a behavior prediction device based on artificial intelligence, which adopts the following technical solutions:

a first acquisition module, used to acquire user behavior data of the target user within a preset time period;

a first generating module, configured to perform feature construction processing on the user behavior data, and generate refueling behavior features based on the user behavior data; wherein, the number of refueling behavior features includes multiple;

a processing module, configured to perform a screening process on the refueling behavior feature based on a principal component analysis algorithm to obtain a first feature, and perform a screening process on the refueling behavior feature based on a preset feature selection algorithm to obtain a second feature;

a first determining module, configured to determine a target refueling behavior feature corresponding to the user behavior data based on the first feature and the second feature;

The second generating module is configured to input the target fueling behavior feature into a preset prediction model, perform prediction processing on the target fueling feature through the prediction model, and generate a fueling behavior prediction result corresponding to the target user .

In order to solve the above-mentioned technical problems, the embodiment of the present application also provides a computer device, which adopts the following technical solutions:

Obtain user behavior data of target users within a preset time period;

Perform feature construction processing on the user behavior data, and generate fueling behavior features based on the user behavior data; wherein, the number of the fueling behavior features includes multiple;

The first feature is obtained by screening the refueling behavior feature based on a principal component analysis algorithm, and the second feature is obtained by screening the refueling behavior feature based on a preset feature selection algorithm;

determining a target fueling behavior feature corresponding to the user behavior data based on the first feature and the second feature;

The target refueling behavior feature is input into a preset prediction model, and prediction processing is performed on the target refueling feature through the prediction model to generate a refueling behavior prediction result corresponding to the target user.

In order to solve the above technical problems, the embodiments of the present application also provide a computer-readable storage medium, which adopts the following technical solutions:

Obtain user behavior data of target users within a preset time period;

Perform feature construction processing on the user behavior data, and generate fueling behavior features based on the user behavior data; wherein, the number of the fueling behavior features includes multiple;

The first feature is obtained by screening the refueling behavior feature based on a principal component analysis algorithm, and the second feature is obtained by screening the refueling behavior feature based on a preset feature selection algorithm;

determining a target fueling behavior feature corresponding to the user behavior data based on the first feature and the second feature;

The target refueling behavior feature is input into a preset prediction model, and prediction processing is performed on the target refueling feature through the prediction model to generate a refueling behavior prediction result corresponding to the target user.

Compared with the prior art, the embodiments of the present application mainly have the following beneficial effects:

After obtaining the user behavior data of the target user within the preset time period, the user behavior data will be constructed and processed first, and the refueling behavior features will be generated based on the user behavior data, and then the refueling behavior features will be screened based on the principal component analysis algorithm. The first feature is obtained, and the refueling behavior feature is screened based on the preset feature selection algorithm to obtain the second feature, and then the target refueling behavior feature corresponding to the user behavior data is determined based on the first feature and the second feature, and finally the target refueling behavior feature is determined. The refueling behavior feature is input into a preset prediction model, and the target refueling feature is predicted through the prediction model to generate a refueling behavior prediction result corresponding to the target user. In the present application, by using the prediction model to predict and process the target refueling behavior characteristics corresponding to the target user, the prediction result of the refueling behavior corresponding to the target user can be quickly and accurately generated, and the processing efficiency and accuracy of the prediction of the refueling behavior intention of the target user can be effectively improved. sex.

Description of drawings

In order to illustrate the solutions in the present application more clearly, the following will briefly introduce the accompanying drawings used in the description of the embodiments of the present application. For those of ordinary skill, other drawings can also be obtained from these drawings without any creative effort.

FIG. 1 is an exemplary system architecture diagram to which the present application can be applied;

2 is a flowchart of an embodiment of the artificial intelligence-based behavior prediction method according to the present application;

3 is a schematic structural diagram of an embodiment of an artificial intelligence-based behavior prediction device according to the present application;

FIG. 4 is a schematic structural diagram of an embodiment of a computer device according to the present application.

Detailed ways

Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the technical field of this application; the terms used herein in the specification of the application are for the purpose of describing specific embodiments only It is not intended to limit the application; the terms "comprising" and "having" and any variations thereof in the description and claims of this application and the above description of the drawings are intended to cover non-exclusive inclusion. The terms "first", "second" and the like in the description and claims of the present application or the above drawings are used to distinguish different objects, rather than to describe a specific order.

Reference herein to an "embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the present application. The appearances of the phrase in various places in the specification are not necessarily all referring to the same embodiment, nor a separate or alternative embodiment that is mutually exclusive of other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

In order to make those skilled in the art better understand the solutions of the present application, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the accompanying drawings.

As shown in FIG. 1 , the system architecture 100 may include terminal devices 101 , 102 , and 103 , a network 104 and a server 105 . The network 104 is a medium used to provide a communication link between the terminal devices 101 , 102 , 103 and the server 105 . The network 104 may include various connection types, such as wired, wireless communication links, or fiber optic cables, among others.

The user can use the terminal devices 101, 102, 103 to interact with the server 105 through the network 104 to receive or send messages and the like. Various communication client applications may be installed on the terminal devices 101 , 102 and 103 , such as web browser applications, shopping applications, search applications, instant messaging tools, email clients, social platform software, and the like.

The terminal devices 101, 102, and 103 may be various electronic devices that have a display screen and support web browsing, including but not limited to smart phones, tablet computers, e-book readers, MP3 players (Moving Picture Experts Group Audio Layer III, moving picture experts). Compression Standard Audio Layer 3), MP4 (Moving PictureExperts Group Audio Layer IV, Moving Picture Experts Compression Standard Audio Layer 4) Players, Laptops and Desktops, etc.

The server 105 may be a server that provides various services, such as a background server that provides support for the pages displayed on the terminal devices 101 , 102 , and 103 .

It should be noted that the artificial intelligence-based behavior prediction method provided in the embodiments of the present application is generally executed by a server/terminal device, and accordingly, an artificial intelligence-based behavior prediction apparatus is generally set in the server/terminal device.

It should be understood that the numbers of terminal devices, networks and servers in FIG. 1 are merely illustrative. There can be any number of terminal devices, networks and servers according to implementation needs.

Continuing to refer to FIG. 2 , a flowchart of one embodiment of the artificial intelligence-based behavior prediction method according to the present application is shown. The artificial intelligence-based behavior prediction method includes the following steps:

Step S201, acquiring user behavior data of a target user within a preset time period.

In this embodiment, the electronic device (for example, the server/terminal device shown in FIG. 1 ) on which the artificial intelligence-based behavior prediction method runs. It should be pointed out that the above wireless connection methods may include, but are not limited to, 3G/4G/5G connection, WiFi connection, Bluetooth connection, WiMAX connection, Zigbee connection, UWB (ultra wideband) connection, and other wireless connections now known or developed in the future. connection method.

In this embodiment, the user behavior data of the target user within a preset time period can be obtained by querying the user's refueling service data in the refueling service webpage related to the user's refueling behavior. The preset time period is not specifically limited, and can be set according to actual usage requirements, for example, it can be set as the first half year from the current time. Specifically, the process of acquiring the user behavior data of the target user within the preset time period may include: first acquiring the page user behavior data within the preset time period, and acquiring the user information of the target user; The user behavior data of the designated page corresponding to the user information is filtered out; the user behavior data of the designated page is subsequently preprocessed to obtain the user behavior data corresponding to the target user.

Step S202: Perform feature construction on the user behavior data, and generate fueling behavior features based on the user behavior data; wherein the number of fueling behavior features includes multiple.

In this embodiment, the user behavior data can be characterized by feature engineering to generate corresponding fueling behavior features. The refueling behavior features at least include user basic features, refueling sequence behavior features, and cumulative behavior features. User basic characteristics: describe the user's personal information, mainly including user ID, user age, user membership level, user type (new user, old user), etc.; refueling sequence behavior characteristics: indicating the statistical analysis of the user's behavior over a period of time, At least include the user's average refueling time interval, the number of days in the observation window from the last refueling to the inspection time window, and the number of times/days the user has performed various behaviors in the observation window, including clicks, browsing, coupons, and placing orders; cumulative behavior characteristics : Represents the statistical analysis of the user's behavior in the past n days, including at least the conversion rate of the user's various behaviors in the n days before the inspection time window. Since the feature extraction of the obtained user behavior data takes into account both the time series behavior characteristics and the cumulative behavior characteristics, the diversity of the data dimension is effectively increased, so that the model prediction processing based on the characteristics of the refueling behavior can ensure the subsequent generation. The accuracy of the prediction results of the target user's refueling behavior.

Step S203 , screening the refueling behavior feature based on a principal component analysis algorithm to obtain a first feature, and performing a screening process on the refueling behavior feature based on a preset feature selection algorithm to obtain a second feature.

In this embodiment, for the above-mentioned specific implementation process of screening the refueling behavior feature based on the principal component analysis algorithm to obtain the first feature, the present application will describe this in further detail in the subsequent specific embodiments. This will not go into too much detail. In addition, the above feature selection algorithm can be determined according to actual requirements, and is preferably an algorithm for removing redundant features from all original features included in the data set. Specifically, the feature selection algorithm is the mRMR (Max-Relevance and Min-Redundancy) algorithm. By using the mRMR algorithm to extract features of all fueling behavior features, the model input and model output can be most correlated. , the redundancy between the model inputs is minimal, so that the subsequent prediction model can have better prediction accuracy when performing prediction processing on the characteristics of the target refueling behavior.

Step S204, determining a target fueling behavior feature corresponding to the user behavior data based on the first feature and the second feature.

In this embodiment, the process of generating the target refueling behavior feature may specifically include: firstly combining the first feature and the second feature to obtain a corresponding feature set; and then performing matching processing on all the features in the feature set, Obtain the third feature that appears repeatedly in the feature set, and use the obtained third feature as the target refueling behavior feature. In this embodiment, the obtained first feature and the second feature are combined and matched to generate the required target refueling behavior feature, because the obtained target refueling behavior feature retains the feature that is more relevant to the refueling prediction process, so It can effectively delete the noise data in the feature space, reduce the saturation recognition modeling error, and reduce the interference effect of noise on the model. And it can also reduce the complexity of the algorithm. Due to the reduction of the dimension of the input data, it is beneficial to reduce the time required for the subsequent prediction model to perform prediction processing, and effectively improve the prediction processing efficiency of the prediction model.

Step S205, inputting the target refueling behavior feature into a preset prediction model, and performing prediction processing on the target refueling feature through the prediction model to generate a refueling behavior prediction result corresponding to the target user.

In this embodiment, the prediction model is generated by training a pre-built LSTM network model based on pre-collected fueling sample data. Using the prediction model can intelligently perform prediction processing on the input target refueling behavior characteristics corresponding to the user behavior data, so as to obtain the refueling behavior prediction result corresponding to the target user. In addition, the refueling behavior prediction result may include having the refueling intention or not having the refueling intention. Further, when it is predicted that the target user has the intention to refuel through the prediction model, the real-time refueling information and user communication information will be intelligently obtained, and the refueling information will be pushed to the target user based on the user communication information, so that the target user can know in time. The current refueling information improves the experience of target users.

After obtaining the user behavior data of the target user within a preset time period, the application will first perform feature construction processing on the user behavior data, generate the refueling behavior feature based on the user behavior data, and then perform the refueling behavior feature based on the principal component analysis algorithm. The screening process obtains the first feature, and the refueling behavior feature is screened based on the preset feature selection algorithm to obtain the second feature, and then the target refueling behavior feature corresponding to the user behavior data is determined based on the first feature and the second feature, and finally The target refueling behavior feature is input into a preset prediction model, and the target refueling feature is predicted through the prediction model to generate a refueling behavior prediction result corresponding to the target user. In the present application, by using the prediction model to predict and process the target refueling behavior characteristics corresponding to the target user, the prediction result of the refueling behavior corresponding to the target user can be quickly and accurately generated, and the processing efficiency and accuracy of the prediction of the refueling behavior intention of the target user can be effectively improved. sex.

In some optional implementations, in step S203, the first feature is obtained by screening the refueling behavior feature based on a principal component analysis algorithm, including the following steps:

A correlation coefficient matrix is constructed based on the refueling behavior characteristics.

In this embodiment, the characteristic equation for constructing the generated correlation coefficient matrix is |R-aI _p |=0.

Dimension reduction processing is performed on the correlation coefficient matrix based on the principal component analysis algorithm to obtain a plurality of principal components.

In this embodiment, the process of performing dimension reduction processing on the correlation coefficient matrix by using the principal component analysis algorithm to select a plurality of corresponding principal components includes: solving the characteristic equation of the above correlation coefficient matrix |R-aI _p |=0, Calculate the characteristic root a, then substitute the obtained characteristic root a into the equation |R-aI _p |x=0, and calculate the eigenvector x, where R represents the correlation coefficient matrix, I _p represents the identity matrix, and each eigenvector x represents a principal component.

Calculate the variance contribution rate of each of the principal components, sort all the variance contribution rates in descending order, and accumulate them in sequence until a preset cumulative variance contribution rate threshold is exceeded.

In this embodiment, the variance contribution rate is a ratio obtained by dividing the eigenvalue of a certain eigenvector by the sum of the eigenvalues of all eigenvectors, and the variance contribution rate represents the ratio of the amount of information contained in this dimension. In addition, the value of the above-mentioned threshold value of the cumulative variance contribution rate is not specifically limited, and can be set according to actual needs, for example, it can be set to 0.99.

A designated principal component corresponding to the accumulation process is selected from all the variance contribution rates, and the feature corresponding to the designated principal component is used as the first feature.

In this embodiment, after all variance contribution rates are sorted in descending order, it is found that the sum of the top five principal components is greater than the cumulative variance contribution rate threshold for the first time through cumulative calculation. , the features corresponding to the top 5 principal components of the ranking will be used as the first features.

By using the principal component analysis method, the present application can quickly screen out the first feature from all the constructed fueling behavior features, which is conducive to the subsequent generation of the final target fueling behavior feature according to the first feature, effectively The accuracy and rationality of the generated target refueling behavior features are guaranteed.

In some optional implementations of this embodiment, step S204 includes the following steps:

The first feature and the second feature are combined to obtain a corresponding feature set.

In this embodiment, the merging process refers to merging and summarizing the first feature and the second feature, so as to obtain a feature set including the first feature and the second feature.

Perform matching processing on all the features in the feature set to obtain a third feature that appears repeatedly in the feature set.

In this embodiment, the matching process may refer to the similarity calculation process. By calculating the similarity between any two features, if the calculated similarity is greater than a preset similarity threshold, it is determined that the two features match each other. feature, that is, the two features belong to the same feature that appears repeatedly in the feature set. In addition, the value of the similarity threshold is not limited, and can be set according to actual requirements.

The third feature is used as the target fueling behavior feature.

In the present application, after using the principal component analysis algorithm to screen the refueling behavior features to obtain the first feature, and using the preset feature selection algorithm to screen the refueling behavior features to obtain the second feature, the first feature and the second feature are analyzed by comparing the first feature and the second feature. Further merging and matching processing is performed to finally determine the target refueling behavior feature corresponding to the user behavior data, which effectively ensures the accuracy and validity of the generated target refueling behavior feature.

In some optional implementations, the prediction model includes an input layer, an LSTM layer, a fully connected layer, and an output layer, and step S205 includes the following steps:

The target fueling behavior feature is input into the input layer of the prediction model, and an output matrix corresponding to the target fueling behavior feature is generated through the input layer.

In this embodiment, the first layer of the prediction model is the input layer, and the input layer is used for data diversity and format conversion. The output matrix generated by the input layer is an output matrix in a preset format obtained by reconstructing the characteristics of the target fueling behavior. The preset format can be set according to actual requirements, such as a 3D format (samples, time_steps, input_dim), where samples represents the number of samples, time_steps represents the time step, and input_dim represents the dimension on each time step.

The output matrix is input into the LSTM layer of the prediction model, and an output vector corresponding to the output matrix is generated through the LSTM layer.

In this embodiment, the LSTM layer controls the information flowing through the unit through a gate mechanism, and the gate mechanism includes an input gate, a forget gate, and an output gate. Among them, the input gate is used to determine the information that needs to be updated; the forget gate is used to control discarding/continue to save the information at the previous moment from the storage state, that is, the storage state is updated through the input gate and the forget gate; the output gate is used to determine the output in the storage state information.

The output vector is input into a fully connected layer of the prediction model, and a prediction classification result corresponding to the output vector is generated based on a preset activation function in the fully connected layer.

In this embodiment, each node of the fully connected layer is connected to all nodes of the previous layer, and the output through the fully connected layer is: x _i =f(∑w _ij *g _j +b _i ), where x _i is the output of the ith neuron in the fully connected layer, _{gj is the jth neuron in the previous layer, w ij is} the connection weight, b _i is the threshold of the ith neuron in the fully connected layer, and f is the activation function. Specifically, the softmax function can be used as the activation function of the fully connected layer. The softmax function can normalize the output values, convert all the output values into probabilities, and the sum of all the probability values is equal to 1.

The prediction classification result is input into the output layer of the prediction model, the prediction classification result is output through the output layer, and the prediction classification result is used as the refueling behavior prediction result corresponding to the target user.

The present application can perform corresponding prediction processing based on the input target refueling behavior characteristics by using the prediction model generated by training, so as to realize the rapid and accurate generation of the refueling behavior prediction result corresponding to the target user, and effectively improve the refueling behavior intention prediction for the target user. processing intelligence.

In some optional implementations, before step S205, the above-mentioned electronic device may further perform the following steps:

Obtain pre-collected fueling sample data.

In this embodiment, the refueling sample data can be obtained by crawling the refueling service data of the user in the relevant refueling service webpage.

The refueling sample data is divided into a training set and a test set according to a preset ratio.

In this embodiment, a preset proportion of data may be randomly selected from the fueling sample data as a training set, and other data remaining in the fueling sample data may be used as a test set. In addition, the above preset ratio is not specifically limited, and can be set according to actual needs. For example, the preset ratio can be set to 70%, and 70% of the data can be randomly selected from the above refueling sample data as a training set. The remaining 30% of the above refueling sample data is used as the test set. In addition, before the model training process of the initial prediction model based on the training set and the test set, the corresponding feature extraction processing will be performed on the training set and the test set based on the above-mentioned principal component analysis algorithm and the preset feature selection algorithm. For the process, refer to the aforementioned feature extraction process for refueling behavior features, which will not be elaborated here. In this embodiment, by performing feature extraction processing on the training set and the test set, the input data dimension of the model can be reduced, the training time of the initial prediction model can be reduced, and the complexity of constructing the initial prediction model can also be reduced, which effectively improves the initial Predictive model building efficiency.

Get a pre-built initial prediction model.

In this embodiment, an initial prediction model is pre-defined and constructed, and the initial prediction model is an LSTM network model composed of an input layer, an LSTM layer, a fully connected layer and an output layer. By setting the internal parameters of the grid, and adjusting the internal parameters of the network . After the data collected by the LSTM network model is converted into the format of the input layer, it is transmitted to the LSTM layer for LSTM network training. After the fully connected layer, the output layer outputs the classification results. In addition, adjust the internal parameters of the network, adjust the learning rate and the number of iterations according to the set value, obtain the input weight, cycle weight and deviation, and use the gate activation function and state activation function to adjust the corresponding incoming gate, forget gate, selected gate and Parameters for the output gate.

Based on the stochastic gradient descent algorithm, the initial prediction model is trained using the training set to obtain a trained initial prediction model.

In this embodiment, the above-mentioned training set includes a plurality of training data and classification label information corresponding to each training data. By using the training data as the input of the initial prediction model, and the classification label information corresponding to each training data as the output of the initial prediction model, the stochastic gradient descent algorithm is used to continuously train the above-mentioned initial prediction model, so that the loss function reaches the expected value. , so that an initial predictive model after training can be generated. In addition, the stochastic gradient descent method is to randomly sample some training data to replace the entire training set. If the sample size is large (such as hundreds of thousands), then only tens of thousands or thousands of samples may be used, and it has been iterated. When the optimal solution is reached, the training speed of the model can be effectively improved. And, for the training process of the initial prediction model using the above-mentioned stochastic gradient descent method, reference may be made to the existing model training process, which will not be repeated here.

The trained initial prediction model is validated using the test set.

In this embodiment, the test set can be input into the initial prediction model after training, and the prediction accuracy rate corresponding to the test set can be calculated. If the prediction accuracy rate is greater than the preset accuracy rate threshold, it is determined that the training initial The prediction model passes the validation, and if the prediction accuracy is not greater than the accuracy threshold, it is determined that the trained initial prediction model fails the validation. In addition, the value of the above-mentioned accuracy threshold is not specifically limited, and can be set according to actual requirements.

If the trained initial prediction model passes the verification, the trained initial prediction model is used as the prediction model.

In this embodiment, after the prediction model is obtained, the prediction model can also be stored in the blockchain network, so that the above-mentioned prediction model generated by training can be stored and managed by using the blockchain, so as to effectively ensure the above-mentioned prediction model. Security and Immutability of Predictive Models.

The present application uses the pre-collected fueling sample data to quickly train and generate the required prediction model based on the stochastic gradient descent algorithm, so that the prediction model can be used to intelligently perform the inputted target fueling behavior characteristics corresponding to the user behavior data. Prediction processing, thereby obtaining the prediction result of the refueling behavior corresponding to the target user, and improving the processing efficiency and processing intelligence of the refueling prediction for the target user.

In some optional implementations of this embodiment, the above-mentioned use of the test set to verify the initial prediction model after training includes the following steps:

Acquire the test set; wherein, the test set includes a plurality of test data and classification information corresponding to each of the test data respectively.

In this embodiment, the above-mentioned test set may be generated based on the above-mentioned refueling sample data set.

The test data is input into the initial prediction model after training, and the first prediction classification result output by the initial prediction model after training and corresponding to each of the test data is obtained.

Based on the classification information, a correctly predicted second predicted classification result among all the first predicted classification results is determined.

In this embodiment, the correctly predicted second predicted classification result refers to the predicted classification result that is the same as the classification information of the corresponding test data among all the first predicted classification results.

A quotient of the second predicted classification result and the first predicted classification result is calculated.

It is judged whether the quotient value is greater than a preset threshold.

In this embodiment, the value of the above-mentioned preset threshold is not specifically limited, and can be set according to actual needs.

If it is greater than the preset threshold, it is determined that the trained initial prediction model has passed the verification; otherwise, it is determined that the trained initial prediction model has not passed the verification.

After obtaining the initial prediction model after training, the application will further use the test set to verify the initial prediction model after training, and only when the initial prediction model after training passes the verification, will the training The initial prediction model of , as the prediction model, ensures the accuracy of the obtained prediction model. And in the future, the prediction model can be used to intelligently predict and process the input target refueling behavior characteristics corresponding to the user behavior data, so as to obtain the refueling behavior prediction result corresponding to the target user, and improve the processing efficiency of the refueling prediction for the target user. Handling intelligence.

In some optional implementations of this embodiment, step S201 includes the following steps:

Acquire page user behavior data within the preset time period, and acquire user information of the target user.

In this embodiment, the page user behavior data may refer to the user behavior data of all users contained in the refueling service webpage. User information may include the user name or user id of the target user.

Based on the user information, the user behavior data of the specified page corresponding to the user information is filtered out from the page user behavior data.

In this embodiment, according to the user information, the user behavior data of the target user that matches the user information, that is, the user behavior data of the specified page, can be filtered from the obtained page user behavior data.

The user behavior data of the specified page is preprocessed to obtain the user behavior data.

In this embodiment, the preprocessing includes data cleaning and data normalization. Among them, when data cleaning is performed on the obtained user behavior data of the specified page, the dropna method in the pandas package in python can be used to remove erroneous data and null values caused by format or collection errors, sort by date, and complete the data cleaning process to obtain Processed user behavior data. At the same time, since in the initial specified page user behavior data, the numerical values and dimensions of different input variables are different, so the data needs to be normalized before subsequent network training, that is, the processed user behavior data All data in are converted into the range of [0,1], the specific function form is: x=x-xmin/xmax-xmin, where xmax is the maximum value of the sample, and xmin is the minimum value of the sample. Further, the above-mentioned preprocessing may also include processing such as filtering noise values and filling in missing values.

The present application can accurately generate the corresponding user behavior data by performing relevant preprocessing on the page user behavior data, which is conducive to the subsequent prediction processing of the user behavior data by using a prediction model, so as to realize the rapid and intelligent generation of the target user's behavior data. Refueling behavior prediction results.

It should be emphasized that, in order to further ensure the privacy and security of the above prediction result of refueling behavior, the above prediction result of refueling behavior can also be stored in a node of a blockchain.

The blockchain referred to in this application is a new application mode of computer technologies such as distributed data storage, point-to-point transmission, consensus mechanism, and encryption algorithm. Blockchain, essentially a decentralized database, is a series of data blocks associated with cryptographic methods. Each data block contains a batch of network transaction information to verify its Validity of information (anti-counterfeiting) and generation of the next block. The blockchain can include the underlying platform of the blockchain, the platform product service layer, and the application service layer.

The embodiments of the present application may acquire and process related data based on artificial intelligence technology. Among them, artificial intelligence (AI) is a theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results. .

The basic technologies of artificial intelligence generally include technologies such as sensors, special artificial intelligence chips, cloud computing, distributed storage, big data processing technology, operation/interaction systems, and mechatronics. Artificial intelligence software technology mainly includes computer vision technology, robotics technology, biometrics technology, speech processing technology, natural language processing technology, and machine learning/deep learning.

Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be implemented by instructing relevant hardware through computer-readable instructions, and the computer-readable instructions can be stored in a computer-readable storage medium. , when the program is executed, it may include the processes of the foregoing method embodiments. The aforementioned storage medium may be a non-volatile storage medium such as a magnetic disk, an optical disk, a read-only memory (Read-Only Memory, ROM), or a random access memory (Random Access Memory, RAM).

It should be understood that although the various steps in the flowchart of the accompanying drawings are sequentially shown in the order indicated by the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless explicitly stated herein, the execution of these steps is not strictly limited to the order and may be performed in other orders. Moreover, at least a part of the steps in the flowchart of the accompanying drawings may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but may be executed at different times, and the execution sequence is also It does not have to be performed sequentially, but may be performed alternately or alternately with other steps or at least a portion of sub-steps or stages of other steps.

Further referring to FIG. 3 , as an implementation of the method shown in FIG. 2 above, the present application provides an embodiment of a behavior prediction device based on artificial intelligence, and the device embodiment corresponds to the method embodiment shown in FIG. 2 , Specifically, the device can be applied to various electronic devices.

As shown in FIG. 3 , the artificial intelligence-based behavior prediction apparatus 300 in this embodiment includes: a first acquisition module 301 , a first generation module 302 , a processing module 303 , a first determination module 304 , and a second generation module 305 . in:

The first obtaining module 301 is used to obtain the user behavior data of the target user within a preset time period;

A first generating module 302, configured to perform feature construction processing on the user behavior data, and generate fueling behavior features based on the user behavior data; wherein, the number of the fueling behavior features includes multiple;

A processing module 303, configured to perform screening processing on the refueling behavior feature based on a principal component analysis algorithm to obtain a first feature, and perform screening processing on the refueling behavior feature based on a preset feature selection algorithm to obtain a second feature;

a first determining module 304, configured to determine a target fueling behavior feature corresponding to the user behavior data based on the first feature and the second feature;

The second generating module 305 is configured to input the target refueling behavior feature into a preset prediction model, perform prediction processing on the target refueling feature through the prediction model, and generate a refueling behavior prediction corresponding to the target user result.

In this embodiment, the operations performed by the above modules or units respectively correspond to the steps of the artificial intelligence-based behavior prediction method in the foregoing embodiment, which will not be repeated here.

In some optional implementations of this embodiment, the processing module 303 includes:

generating a submodule for constructing a correlation coefficient matrix based on the refueling behavior feature;

a first processing submodule, configured to perform dimensionality reduction processing on the correlation coefficient matrix based on the principal component analysis algorithm to obtain a plurality of principal components;

The second processing sub-module is configured to calculate the variance contribution rate of each principal component, sort all the variance contribution rates in descending order, and accumulate them in sequence until exceeding the preset cumulative variance contribution rate threshold ;

The first determination sub-module is configured to screen out a specified principal component corresponding to the accumulation process from all the variance contribution rates, and use the feature corresponding to the specified principal component as the first feature.

In this embodiment, the operations performed by the above modules or units respectively correspond to the steps of the artificial intelligence-based behavior prediction method in the foregoing embodiment, which will not be repeated here.

In some optional implementations of this embodiment, the first determining module 304 includes:

a third processing sub-module, configured to perform merging processing on the first feature and the second feature to obtain a corresponding feature set;

a fourth processing submodule, configured to perform matching processing on all the features in the feature set, and obtain a third feature that appears repeatedly in the feature set;

The second determination submodule is configured to use the third feature as the target refueling behavior feature.

In this embodiment, the operations performed by the above modules or units respectively correspond to the steps of the artificial intelligence-based behavior prediction method in the foregoing embodiment, and details are not described herein again.

In some optional implementations of this embodiment, the prediction model includes an input layer, an LSTM layer, a fully connected layer, and an output layer, and the second generation module 305 includes:

a first generating submodule, configured to input the target fueling behavior feature into an input layer of the prediction model, and generate an output matrix corresponding to the target fueling behavior feature through the input layer;

a second generating submodule, configured to input the output matrix into the LSTM layer of the prediction model, and generate an output vector corresponding to the output matrix through the LSTM layer;

a third generating submodule, configured to input the output vector into a fully connected layer of the prediction model, and generate a prediction classification result corresponding to the output vector based on a preset activation function in the fully connected layer;

The third determination sub-module is configured to input the predicted classification result into the output layer of the prediction model, output the predicted classification result through the output layer, and use the predicted classification result as the target user The corresponding prediction result of refueling behavior.

In this embodiment, the operations performed by the above modules or units respectively correspond to the steps of the artificial intelligence-based behavior prediction method in the foregoing embodiment, which will not be repeated here.

In some optional implementations of this embodiment, the artificial intelligence-based behavior prediction apparatus further includes:

The second acquisition module is used for acquiring pre-collected fueling sample data;

a dividing module, configured to divide the refueling sample data into a training set and a test set according to a preset ratio;

The third acquisition module is used to acquire the pre-built initial prediction model;

A training module for training the initial prediction model by using the training set based on the stochastic gradient descent algorithm to obtain the initial prediction model after training;

a verification module, configured to use the test set to verify the trained initial prediction model;

The second determination module is configured to use the trained initial prediction model as the prediction model if the trained initial prediction model passes the verification.

In this embodiment, the operations performed by the above modules or units respectively correspond to the steps of the artificial intelligence-based behavior prediction method in the foregoing embodiment, which will not be repeated here.

In some optional implementations of this embodiment, the verification module includes:

a first acquisition submodule, configured to acquire the test set; wherein, the test set includes a plurality of test data and classification information corresponding to each of the test data respectively;

The second obtaining sub-module is configured to input the test data into the initial prediction model after training, and obtain the first prediction classification result corresponding to each test data output by the initial prediction model after training ;

a fourth determination submodule, configured to determine, based on the classification information, a second prediction classification result that is correctly predicted among all the first prediction classification results;

a calculation submodule for calculating the quotient of the second predicted classification result and the first predicted classification result;

a judging submodule for judging whether the quotient is greater than a preset threshold;

A determination sub-module, configured to determine that the trained initial prediction model passes the verification if it is greater than the preset threshold, otherwise the trained initial prediction model fails the verification.

In this embodiment, the operations performed by the above modules or units respectively correspond to the steps of the artificial intelligence-based behavior prediction method in the foregoing embodiment, which will not be repeated here.

In some optional implementations of this embodiment, the first obtaining module 301 includes:

a third acquisition sub-module, configured to acquire page user behavior data within the preset time period, and acquire user information of the target user;

A screening submodule, configured to screen out the user behavior data of the specified page corresponding to the user information from the page user behavior data based on the user information;

The fifth processing submodule is used for the submodule, and is used for preprocessing the user behavior data of the specified page to obtain the user behavior data.

In this embodiment, the operations performed by the above modules or units respectively correspond to the steps of the artificial intelligence-based behavior prediction method in the foregoing embodiment, which will not be repeated here.

To solve the above technical problems, the embodiments of the present application also provide computer equipment. Please refer to FIG. 4 for details. FIG. 4 is a block diagram of a basic structure of a computer device according to this embodiment.

The computer device 4 includes a memory 41, a processor 42, and a network interface 43 that communicate with each other through a system bus. It should be noted that only the computer device 4 with components 41-43 is shown in the figure, but it should be understood that it is not required to implement all of the shown components, and more or less components may be implemented instead. Among them, those skilled in the art can understand that the computer device here is a device that can automatically perform numerical calculation and/or information processing according to pre-set or stored instructions, and its hardware includes but is not limited to microprocessors, special-purpose Integrated circuit (ApplicationSpecific Integrated Circuit, ASIC), programmable gate array (Field-Programmable GateArray, FPGA), digital processor (Digital Signal Processor, DSP), embedded equipment, etc.

The computer equipment may be a desktop computer, a notebook computer, a palmtop computer, a cloud server and other computing equipment. The computer device can perform human-computer interaction with the user through a keyboard, a mouse, a remote control, a touch pad or a voice control device.

The memory 41 includes at least one type of readable storage medium, and the readable storage medium includes flash memory, hard disk, multimedia card, card-type memory (for example, SD or DX memory, etc.), random access memory (RAM), static Random Access Memory (SRAM), Read Only Memory (ROM), Electrically Erasable Programmable Read Only Memory (EEPROM), Programmable Read Only Memory (PROM), Magnetic Memory, Magnetic Disk, Optical Disk, etc. In some embodiments, the memory 41 may be an internal storage unit of the computer device 4 , such as a hard disk or a memory of the computer device 4 . In other embodiments, the memory 41 may also be an external storage device of the computer device 4 , such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, flash memory card (FlashCard) and so on. Of course, the memory 41 may also include both the internal storage unit of the computer device 4 and its external storage device. In this embodiment, the memory 41 is generally used to store the operating system and various application software installed on the computer device 4 , such as computer-readable instructions for a behavior prediction method based on artificial intelligence. In addition, the memory 41 can also be used to temporarily store various types of data that have been output or will be output.

The processor 42 may be a central processing unit (Central Processing Unit, CPU), a controller, a microcontroller, a microprocessor, or other data processing chips in some embodiments. This processor 42 is typically used to control the overall operation of the computer device 4 . In this embodiment, the processor 42 is configured to execute computer-readable instructions stored in the memory 41 or process data, for example, computer-readable instructions for executing the artificial intelligence-based behavior prediction method.

The network interface 43 may include a wireless network interface or a wired network interface, and the network interface 43 is generally used to establish a communication connection between the computer device 4 and other electronic devices.

Compared with the prior art, the embodiments of the present application mainly have the following beneficial effects:

In the embodiment of the present application, after acquiring the user behavior data of the target user within the preset time period, the user behavior data will be characterized firstly, and the refueling behavior feature will be generated based on the user behavior data, and then based on the principal component analysis algorithm The refueling behavior feature is screened to obtain the first feature, and the refueling behavior feature is screened based on the preset feature selection algorithm to obtain the second feature, and then the target refueling corresponding to the user behavior data is determined based on the first feature and the second feature. Finally, the target refueling behavior characteristics are input into the preset prediction model, and the target refueling characteristics are predicted and processed through the prediction model, and the refueling behavior prediction results corresponding to the target users are generated. In the present application, by using the prediction model to predict and process the characteristics of the target refueling behavior corresponding to the target user, the prediction result of the refueling behavior corresponding to the target user can be quickly and accurately generated, and the processing efficiency and accuracy of the prediction of the refueling behavior intention of the target user can be effectively improved. sex.

The present application also provides another embodiment, that is, to provide a computer-readable storage medium, where the computer-readable storage medium stores computer-readable instructions, and the computer-readable instructions can be executed by at least one processor to The at least one processor is caused to perform the steps of the artificial intelligence-based behavior prediction method as described above.

Compared with the prior art, the embodiments of the present application mainly have the following beneficial effects:

In the embodiment of the present application, after acquiring the user behavior data of the target user within the preset time period, the user behavior data will be characterized firstly, and the refueling behavior feature will be generated based on the user behavior data, and then based on the principal component analysis algorithm The refueling behavior feature is screened to obtain the first feature, and the refueling behavior feature is screened based on the preset feature selection algorithm to obtain the second feature, and then the target refueling corresponding to the user behavior data is determined based on the first feature and the second feature. Finally, the target refueling behavior characteristics are input into the preset prediction model, and the target refueling characteristics are predicted and processed through the prediction model, and the refueling behavior prediction results corresponding to the target users are generated. In the present application, by using the prediction model to predict and process the characteristics of the target refueling behavior corresponding to the target user, the prediction result of the refueling behavior corresponding to the target user can be quickly and accurately generated, and the processing efficiency and accuracy of the prediction of the refueling behavior intention of the target user can be effectively improved. sex.

From the description of the above embodiments, those skilled in the art can clearly understand that the method of the above embodiment can be implemented by means of software plus a necessary general hardware platform, and of course can also be implemented by hardware, but in many cases the former is better implementation. Based on this understanding, the technical solution of the present application can be embodied in the form of a software product in essence or in a part that contributes to the prior art, and the computer software product is stored in a storage medium (such as ROM/RAM, magnetic disk, CD-ROM), including several instructions to make a terminal device (which may be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) execute the methods described in the various embodiments of this application.

Obviously, the above-described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments. The accompanying drawings show the preferred embodiments of the present application, but do not limit the scope of the patent of the present application. This application may be embodied in many different forms, rather these embodiments are provided so that a thorough and complete understanding of the disclosure of this application is provided. Although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing specific embodiments, or perform equivalent replacements for some of the technical features. . Any equivalent structure made by using the contents of the description and drawings of the present application, which is directly or indirectly used in other related technical fields, is also within the scope of protection of the patent of the present application.

Claims (10)

1. a kind of behavior prediction method based on artificial intelligence, is characterized in that, comprises the following steps: Obtain user behavior data of target users within a preset time period; Perform feature construction processing on the user behavior data, and generate fueling behavior features based on the user behavior data; wherein, the number of the fueling behavior features includes multiple; The first feature is obtained by screening the refueling behavior feature based on a principal component analysis algorithm, and the second feature is obtained by screening the refueling behavior feature based on a preset feature selection algorithm; determining a target fueling behavior feature corresponding to the user behavior data based on the first feature and the second feature; The target refueling behavior feature is input into a preset prediction model, and prediction processing is performed on the target refueling feature through the prediction model to generate a refueling behavior prediction result corresponding to the target user. 2. The behavior prediction method based on artificial intelligence according to claim 1, is characterized in that, the step that the described refueling behavior feature is screened and processed based on principal component analysis algorithm to obtain the first feature, specifically comprises: constructing a correlation coefficient matrix based on the refueling behavior feature; Perform dimensionality reduction processing on the correlation coefficient matrix based on the principal component analysis algorithm to obtain a plurality of principal components; Calculate the variance contribution rate of each of the principal components, sort all the variance contribution rates in descending order, and accumulate them in turn until exceeding the preset cumulative variance contribution rate threshold; A designated principal component corresponding to the accumulation process is selected from all the variance contribution rates, and the feature corresponding to the designated principal component is used as the first feature. 3 . The behavior prediction method based on artificial intelligence according to claim 1 , wherein the method of determining the target refueling behavior feature corresponding to the user behavior data based on the first feature and the second feature. 4 . steps, including: Merging the first feature and the second feature to obtain a corresponding feature set; Perform matching processing on all the features in the feature set, and obtain a third feature that appears repeatedly in the feature set; The third feature is used as the target fueling behavior feature. 4. The behavior prediction method based on artificial intelligence according to claim 1, wherein the prediction model comprises an input layer, an LSTM layer, a fully connected layer and an output layer, and the target refueling behavior feature is input into a In the preset prediction model, the prediction process is performed on the target fueling feature through the prediction model, and the steps of generating a fueling behavior prediction result corresponding to the target user specifically include: Inputting the target refueling behavior feature into an input layer of the prediction model, and generating an output matrix corresponding to the target refueling behavior feature through the input layer; Inputting the output matrix into the LSTM layer of the prediction model, and generating an output vector corresponding to the output matrix through the LSTM layer; inputting the output vector into a fully connected layer of the prediction model, and generating a prediction classification result corresponding to the output vector based on a preset activation function in the fully connected layer; The prediction classification result is input into the output layer of the prediction model, the prediction classification result is output through the output layer, and the prediction classification result is used as the refueling behavior prediction result corresponding to the target user. 5. The behavior prediction method based on artificial intelligence according to claim 1, characterized in that, in the described target refueling behavior feature input into a preset prediction model, the target refueling feature is analyzed by the prediction model. Before the step of performing prediction processing and generating the prediction result of the refueling behavior corresponding to the target user, the method further includes: Obtain pre-collected fueling sample data; Divide the refueling sample data into a training set and a test set according to a preset ratio; Get a pre-built initial prediction model; Based on the stochastic gradient descent algorithm, use the training set to train the initial prediction model to obtain a trained initial prediction model; Validating the trained initial prediction model using the test set; If the trained initial prediction model passes the verification, the trained initial prediction model is used as the prediction model. 6. The artificial intelligence-based behavior prediction method according to claim 5, wherein the step of using the test set to verify the initial prediction model after the training, and judging whether the verification passes, specifically includes : obtaining the test set; wherein, the test set includes a plurality of test data, and classification information corresponding to each of the test data respectively; Inputting the test data into the initial prediction model after training, and obtaining the first prediction classification result output by the initial prediction model after training and corresponding to each of the test data respectively; determining, based on the classification information, a second prediction classification result that is correctly predicted among all the first prediction classification results; calculating the quotient of the second predicted classification result and the first predicted classification result; judging whether the quotient value is greater than a preset threshold; If it is greater than the preset threshold, it is determined that the trained initial prediction model has passed the verification; otherwise, it is determined that the trained initial prediction model has not passed the verification. 7. The artificial intelligence-based behavior prediction method according to claim 1, wherein the step of obtaining the user behavior data of the target user in a preset time period specifically includes: Acquiring page user behavior data within the preset time period, and acquiring user information of the target user; Based on the user information, filter out the user behavior data of the specified page corresponding to the user information from the page user behavior data; The user behavior data of the specified page is preprocessed to obtain the user behavior data. 8. A behavior prediction device based on artificial intelligence, characterized in that, comprising: a first acquisition module, used to acquire user behavior data of the target user within a preset time period; a first generating module, configured to perform feature construction processing on the user behavior data, and generate refueling behavior features based on the user behavior data; wherein, the number of refueling behavior features includes multiple; a processing module, configured to perform a screening process on the refueling behavior feature based on a principal component analysis algorithm to obtain a first feature, and perform a screening process on the refueling behavior feature based on a preset feature selection algorithm to obtain a second feature; a first determining module, configured to determine a target refueling behavior feature corresponding to the user behavior data based on the first feature and the second feature; The second generation module is configured to input the target fueling behavior feature into a preset prediction model, perform prediction processing on the target fueling feature through the prediction model, and generate a fueling behavior prediction result corresponding to the target user . 9. A computer device comprising a memory and a processor, wherein computer-readable instructions are stored in the memory, the processor implementing the computer-readable instructions as claimed in any one of claims 1 to 7 when the processor executes the computer-readable instructions Steps of an artificial intelligence-based approach to behavior prediction. 10. A computer-readable storage medium, wherein computer-readable instructions are stored on the computer-readable storage medium, and when the computer-readable instructions are executed by a processor, any one of claims 1 to 7 is implemented. The steps of the artificial intelligence-based behavior prediction method described in item.

Images