TWI849270B - Article recommendation method and system which presents articles in map - Google Patents

Abstract

An article recommendation method is executed on an article recommendation server, comprising: obtaining a plurality of articles; performing article pre-processing on the obtained plurality of articles to obtain pre-processing data of each article; establishing an article model of each article according to the pre-processing data of each article; calculating article similarities of the plurality of articles, and obtaining a plurality of recommended articles accordingly; classifying the plurality of recommended articles according to a plurality of thinking words, and obtaining a map location of each recommended article; and presenting the plurality of recommended articles on an article recommendation page in a map presentation manner according to their map locations.

Description

Article recommendation method and system for presenting articles in the form of maps

The present invention relates to an article recommendation method and system, and in particular to an article recommendation method and system for presenting articles in the form of a map.

Article recommendation systems or methods can recommend articles related to the article currently being viewed by the user, or recommend articles related to the keywords input by the user, so that the user can read the related articles, thereby allowing the user to learn specific knowledge or concepts. However, the article recommendation systems or methods of the prior art cannot allow the user to clearly know the relationship between the article or keyword being read and the related articles. Therefore, there is a need to provide an article recommendation method and system that can allow the user to know the relationship between the recommended article and the article or keyword being viewed.

Based on one of the objectives to be achieved and the technical problems to be solved by the present invention, an embodiment of the present invention provides an article recommendation method, which is executed on an article recommendation server, including: obtaining multiple articles; performing article pre-processing on the obtained multiple articles to obtain pre-processing data of each article; establishing an article model of each article based on the pre-processing data of each article; calculating the article similarity of the multiple articles, and obtaining multiple recommended articles accordingly; classifying the multiple recommended articles according to multiple thinking words, and obtaining the map location of each recommended article; and presenting the multiple recommended articles on the article recommendation page in a map presentation manner according to their map locations.

According to the aforementioned technical features, the plurality of articles are obtained by searching based on keywords.

According to the aforementioned technical features, the plurality of articles are obtained based on the article that the user is currently reading.

According to the aforementioned technical features, the multiple thinking words are "industry thinking", "business thinking" and "investment thinking".

According to the aforementioned technical features, the article recommendation method further includes: obtaining a recommended series of articles and a recommended financial product, wherein the recommended series of articles and the recommended financial product are presented on the article recommendation page.

According to the aforementioned technical features, the article recommendation page further has a search box, and the keyword field of the search box is a keyword related to the article, so that the user can click to search.

According to the aforementioned technical features, the article model is the vector coordinates of the pre-processed data converted based on the article-vector model.

According to the aforementioned technical features, the similarities of the plurality of articles are calculated, and the first few articles with higher similarities among the plurality of articles are used as the plurality of recommended articles.

Based on one of the objectives to be achieved and the technical problems to be solved by the present invention, an embodiment of the present invention provides an article recommendation system, including: a user device; and an article recommendation server, which is communicatively connected to the user device, and the article recommendation server is used to: obtain multiple articles; perform article pre-processing on the obtained multiple articles to obtain pre-processing data of each article; establish an article model of each article based on the pre-processing data of each article; calculate the article similarity of the multiple articles, and obtain multiple recommended articles accordingly; classify the multiple recommended articles according to multiple thinking words, and obtain the map location of each recommended article; and present the multiple recommended articles on the article recommendation page in a map presentation manner according to their map locations.

In short, the embodiment of the present invention provides an article recommendation method and system that allows users to know the relationship between recommended articles and articles or keywords currently being viewed, wherein articles are presented in the form of maps so that users can know the relationship between articles, making it easier for users to read related articles and learn related knowledge.

In order to fully understand the purpose, features and effects of the present invention, the present invention is described in detail through the following specific embodiments and the accompanying drawings.

The present invention provides an article recommendation system and method, wherein the article recommendation page displayed on a mobile phone screen or a computer screen presents multiple articles in a map, so that the user can know the relationship between the article or keyword being read and other articles. The article recommendation system includes at least one user device and an article recommendation server, wherein the user device and the article recommendation server are connected to each other in communication, the article recommendation server can be implemented by hardware circuits and software, or all by hardware circuits, the user device can be a smart phone, a laptop, a desktop computer, a tablet computer or a netbook, and the present invention is not limited thereto.

In one embodiment of the invention, a user can send input keywords to an article recommendation server through a user device, so that the article recommendation server obtains multiple articles based on the keywords, wherein the multiple articles calculate the association values between the articles and the thought words based on their contents to obtain the map locations of the articles, wherein the map locations of the articles include the association values of each thought word and its associated words with the articles. Then, the article recommendation page sent by the article recommendation server to the user device for presentation will present each article to the user in the form of a map based on the map locations of each article.

In another embodiment of the present invention, the article recommendation server obtains multiple articles related to the article being read by the user according to the article being read by the user, calculates the map positions of the article being read and the multiple articles according to the above method, and sends the article recommendation page to the user device. The above method of obtaining multiple related articles according to the article being read can be to obtain multiple specific words of the article being read as keywords, and obtain multiple articles according to these keywords.

The aforementioned thinking words may be "business thinking", "industry thinking" and "investment thinking", but the present invention is not limited thereto. The number of thinking words is two or more, and the thinking words and their number may be customized by the service provider according to actual usage. For example, if the provider of the article recommendation server is a provider specializing in semiconductor technology knowledge, the thinking words may be "semiconductor materials", "semiconductor physics" and "semiconductor manufacturing process".

After explaining the inventive concept of the article recommendation system and method of the embodiment of the present invention, the details of the article recommendation system and method of the embodiment of the present invention will be further explained. First, please refer to Figure 1, which is a block diagram of the article recommendation system of the embodiment of the present invention. The article recommendation system 1 includes an article recommendation server 11 and at least one user device 12, wherein the article recommendation server 11 is communicatively connected to the user device 12, and can also be communicatively connected to an external server 13. The article recommendation server 11 can be implemented by hardware circuits and software, or entirely by hardware circuits, and the user device 12 can be a smart phone, a laptop, a desktop computer, a tablet computer or a netbook, and the present invention is not limited thereto.

In one embodiment, the article recommendation server 11 receives a keyword from a user device 12, and searches for multiple articles related to the keyword from the article recommendation server 11 and the external server 13 according to the keyword. Then, the article recommendation server 11 performs article pre-processing on the obtained multiple articles to obtain pre-processing data, for example, performing semantic analysis to find multiple specific words in the article and calculating the frequency of the specific words. Then, the article recommendation server 11 establishes an article model based on the pre-processing data, and the article model is the vector coordinates after the processing data is converted based on the article-vector model.

Afterwards, the article recommendation server 11 uses the article models of the multiple articles to calculate the article similarity of the multiple articles, wherein the article similarity is calculated by calculating the distance between the vector coordinates of the article models of the multiple articles and the vector coordinates corresponding to the keywords. The closer the distance, the higher the article similarity. Since there may be a large number of articles, the first few articles with higher article similarity will be used as recommended articles.

Next, the article recommendation server 11 categorizes the multiple recommended articles according to the thought words. Afterwards, the article recommendation server 11 can further obtain recommended financial products and recommended series of articles, wherein the recommended financial products and the recommended series of articles are related to the keywords. Finally, the article recommendation server 11 merges the recommended financial products, the recommended series of articles, and the multiple recommended articles on the article recommendation page, and provides them to the user device 12 for display on the screen of the user device 12. The multiple recommended articles are presented on the article recommendation page in a map presentation manner according to their map locations.

In another embodiment, the article recommendation server 11 does not receive the keyword of the user device 12, but receives the article selected by the user device 12 to be read, and obtains multiple articles according to the article being read by the user of the user device 12. The above method of obtaining multiple related articles according to the article being read by the user can be to obtain multiple specific words of the article being read as keywords, and obtain multiple articles according to these keywords. Afterwards, the method of calculating the article similarity and classifying multiple recommended articles according to the thinking words is the same as described above.

Next, the implementation of the article recommendation server 11 is further described, but the present invention is not limited to the implementation of the article recommendation server 11 described below. Please refer to FIG. 2, which is a block diagram of the article recommendation server of an embodiment of the present invention. In FIG. 2, the article recommendation server 11 includes a processing unit 21, an input/output interface unit 22, a storage unit 23, and a memory 24, wherein the processing unit 21 is electrically connected to the input/output interface unit 22, the storage unit 23, and the memory 24.

In this embodiment, the processing unit 21 is used to execute the software program to perform article recommendation. The input-output interface unit 22 includes a communication module to enable the article recommendation server 11 to communicate with the user device, thereby receiving the keywords of the user device and transmitting the search results to the user device. The storage unit 23 stores the aforementioned software program, and can also be planned with a specific database unit 231 to store various types of databases. The memory 24 serves as a data storage space when the processing unit 21 executes the software program. It should be noted that in other embodiments, the storage unit 23 may not have a specific part for storing various databases, or various databases may be stored outside the article recommendation server 11. In short, the present invention is not limited by the storage method of various databases.

Next, please refer to Figure 3, which is a flow chart of the article recommendation method of an embodiment of the present invention. The article recommendation method can be executed by the above-mentioned article recommendation server 11, and each step is described as follows. First, in step S31, the article recommendation server 11 obtains multiple articles, wherein the multiple articles can be obtained based on the keywords input by the user device 12 or the articles that the user is reading. Then, in step S32, the multiple articles obtained are pre-processed to obtain pre-processed data. Then, in step S33, an article model is established based on the pre-processed data, wherein the article model is the vector coordinates of the pre-processed data after conversion based on the article-vector model.

Then, in step S34, the article similarity of multiple articles is calculated. Then, in step S35, multiple recommended articles are classified according to thinking words. After that, in step S36, recommended financial products and recommended series of articles are obtained, wherein the recommended financial products and the recommended series of articles are related to keywords. Finally, in step S37, an article recommendation page is provided, wherein multiple recommended articles are presented on the article recommendation page in a map presentation manner according to their map positions. The way in which multiple recommended articles are presented on the article recommendation page in a map presentation manner according to their map positions is to project the vector coordinates of the above-mentioned recommended articles converted from the article-vector model into a two-dimensional space, and distribute them according to different categories of thinking words.

For example, the thinking words are "industry thinking", "business thinking" and "investment thinking", and the vector coordinates of a recommended article converted by the article-vector model are roughly close to the two thinking words "industry thinking" and "business thinking", so the map position of the recommended article on the map is close to the two thinking words "industry thinking" and "business thinking", but not close to "investment thinking".

Next, please refer to FIG. 4A , which is a schematic diagram of an article recommendation method or system of an embodiment of the present invention presenting an article recommendation page on a mobile phone screen. In FIG. 4A , a map 40 on the article recommendation page presents a plurality of recommended articles categorized as “industry thinking”, “business thinking”, and “investment thinking”, and a plurality of articles on the map will be presented at their corresponding map positions. For example, recommended articles 41 and 42 are located at map positions 401 and 402 on the map 40 , respectively. In addition, there is a search box 43 on the article recommendation page, and the keyword field of the search box 43 is a keyword related to the article, so that the user can click to search, so that the article recommendation system or method recommends articles based on the keywords. In addition, the article recommendation page also has a recommended series of articles 44. In addition, please refer to Figure 4B, which is a schematic diagram of the article recommendation method or system of the embodiment of the present invention presenting the article recommendation page on the computer screen. The difference between Figure 4A and Figure 4B is that the layout of the article recommendation page will be different due to the difference in screens.

In summary, the article recommendation method and system for presenting articles in a map according to the embodiments of the present invention can recommend related articles based on keywords or articles that users are viewing, and the recommended articles are presented in a map so that users can know the relationship between articles, making it convenient for users to read related articles and learn related knowledge.

The present invention has been disclosed in the above with preferred embodiments, but those skilled in the art should understand that the above embodiments are only used to describe the present invention and should not be interpreted as limiting the scope of the present invention. It should be noted that all changes and substitutions equivalent to the above embodiments should be included in the scope of the present invention. Therefore, the protection scope of the present invention shall be based on the scope defined by the application patent.

1: Article recommendation system 11: Article recommendation server 12: User device 13: External server 2: Article recommendation server 21: Processing unit 22: Input/output interface unit 23: Storage unit 231: Database unit S31~S37: Steps

FIG1 is a block diagram of an article recommendation system according to an embodiment of the present invention.

FIG2 is a block diagram of an article recommendation server according to an embodiment of the present invention.

FIG3 is a flow chart of the article recommendation method according to an embodiment of the present invention.

FIG4A is a schematic diagram of an article recommendation method or system according to an embodiment of the present invention presenting an article recommendation page on a mobile phone screen.

FIG4B is a schematic diagram of an article recommendation method or system according to an embodiment of the present invention presenting an article recommendation page on a computer screen.

S31~S37: Steps

Claims (3)

An article recommendation method is executed on an article recommendation server, and multiple articles are presented as a map to let users know the relationship between the article or keyword being read and other articles; the method includes: obtaining keywords according to the article being read by the user, searching for multiple articles related to the keyword according to the keyword to obtain the multiple articles; performing article pre-processing on the obtained multiple articles to obtain the article pre-processing data of each article; wherein the pre-processing is to perform semantic analysis to find multiple specific words in the article and calculate the word frequency of the specific words, so as to use the multiple specific words and the word frequency of the specific words as pre-processing data; establish an article model of each article according to the pre-processing data of each article; wherein the article model includes multiple specific words with a word frequency greater than a predetermined frequency and the word frequency of the specific words; use the article models of multiple articles to calculate the calculation The method comprises the steps of calculating article similarity of a plurality of articles, and selecting a plurality of recommended articles from the plurality of articles based on the article similarity; wherein calculating the article similarity is calculating the correlation value between a specific word and a keyword in an article model, and a plurality of articles with article similarity above a predetermined similarity are used as the plurality of recommended articles; classifying the plurality of recommended articles according to a plurality of thinking words, and obtaining a map position of each of the recommended articles; wherein the characteristic value of the article model of each of the recommended articles is The association value between the fixed word and each thought word is calculated to obtain the map position of the recommended article; the multiple recommended articles are presented on the article recommendation page in a map presentation manner according to their map positions; and the recommended series of articles and recommended financial products are obtained, and the recommended series of articles and the recommended financial products are presented on the article recommendation page, wherein the recommended financial products and the recommended series of articles are associated with the keyword. For example, in the article recommendation method of request 1, the multiple thinking words are "industry thinking", "business thinking" and "investment thinking". As in the article recommendation method of claim 1, the article recommendation page further has a search box to allow the user to enter the keyword.