TWI595371B - Legal information search system,? search processing method and computer readable media recorder - Google Patents

Description

Legal data search system, search processing method and computer readable recording medium

The present invention relates to a data search system and a search processing method, and more particularly to a legal data search system, a search processing method and a computer readable recording medium for searching and sorting the keyword groups in a single field.

Generally speaking, the scope of legal materials mainly includes trial-related laws and regulations, interpretations, judgments, legal cases, announcements of ex-rights judgments, etc., and can be used to conduct legal data inquiry through a legal data search system based on the scope of legal information to be inquired. For example, the Court's Jurisprudential Information Retrieval System (available at http://jirs.judicial.gov.tw/Index.htm ). When using the conventional legal data search system, it is usually necessary to first determine the scope of the query, such as querying regulations or judgments, and then querying the regulations or judgments after clicking on the entry page.

Figure 1 is a schematic diagram of the interface used by the conventional legal data search system. As shown in Figure 1, if you want to use the traditional legal data search system to use the interface 1 to make a judgment, the user must first confirm the referee's font size, year, court level and region, and use the legal data search system. 1 The displayed court name, referee category, judgment font number, judgment case, judgment date, referee main text and full-text search terms are entered into the corresponding attribute keywords, and the corresponding legal materials can be found after the search is performed. However, traditional The legal data search system can only search for data that exactly matches the keyword, and the keyword information such as the judgment font size, the court level and the region are not easy for the user to obtain and the grammar is difficult to understand, so the keyword is often input. After the search, the search results of "check no data" are obtained. In other words, the traditional legal data search system is not only complicated to operate, but also cannot provide relevant legal materials that are related to keywords.

For example, when the case is to be "infringed on the property rights of the patent right", if the patent infringement is entered in the field by the judgment, no information can be searched. When searching for the referee's font number as "Sixty-seventh Shang Yi Zi No. 61", if you enter "63 Years of Shang Yi Zi No. 61" in the judgment font size field, you will not be able to find any information. Furthermore, in the above case, if the court of the search is entered incorrectly, no legal information can be found. Since the input keyword must accurately match the content of the legal data, the legal data can be searched, which may result in the inability to retrieve the legal data to be inquired, and the related legal data cannot be retrieved, resulting in inconvenience in searching. And the retrieval process is cumbersome and time-consuming, and cannot provide a user-friendly operation experience.

In addition, the list of results searched by the conventional legal data search system is based on the date order, but it cannot be sorted according to the relevance of the judgment and the keyword. Therefore, the ranking of the search results is often not the same as the previous judgment. The most relevant judgments make the accuracy of the search low, causing users to read and filter a lot, and the time for searching and reading. In addition, in addition to the judgment of the judgment, the same problems exist in the trial of relevant laws and regulations, the retrieval of judgments, the retrieval of legal cases or the retrieval of ex-rights judgments. On the whole, the retrieval accuracy of the traditional legal data search system is low, and the retrieval process is cumbersome and time-consuming and inconvenient.

In view of this, it is necessary to develop a legal data search system and a search processing method to solve the problems faced by the prior art.

The purpose of this case is to provide a legal data search system and a search processing method, which can provide relevant legal materials related to keywords, improve the accuracy of retrieval, provide a friendly and easy-to-operate user experience, and save users. The time of retrieval and reading.

In order to achieve the above objectives, one of the broader aspects of the case provides a legal data search system, which includes: a legal database, storing a plurality of legal data; a client device, including an intelligent search interface, wherein the intelligent search interface is provided to the user. Entering a keyword group; and a search server, connecting and communicating with the legal database and the client device, and including a processor, wherein the processor performs a legal data search operation according to the keyword group of the user device, and the processor includes: a word tag and a selective extension module, which mark one or a plurality of keyword sentences of the keyword group and selectively perform an expansion operation to obtain a plurality of candidate keyword sentences, wherein each candidate keyword sentence respectively There is a mark, and the mark is associated with the category of the candidate keyword sentence; the keyword search field and the weight adjustment module select the corresponding search field of each candidate keyword according to the plurality of candidate keyword sentences and the associated tag and Corresponding weighting parameter; search and relevance score calculation module, based on a plurality of candidate keywords Searching for multiple legal data of the legal database to obtain multiple candidate legal data, and calculating the relevance scores of each candidate legal data for the plurality of legal materials; and sorting module, according to the relevance of the plurality of candidate legal materials The score is subjected to a relevance ranking job, and the plurality of candidate legal data sorted by the relevance degree is provided to the user device.

In order to achieve the above objectives, another broader mode of the present invention is to provide a method for searching for legal data for use in a legal data search system for performing legal data retrieval operations, wherein the legal data search system includes a retrieval server, and the retrieval server includes The processor is configured to perform a law data retrieval operation, and the legal data search processing method comprises the steps of: (a) receiving a keyword group, marking one or a plurality of keyword phrases, and selectively entering the keyword group; Extending the job to obtain a plurality of candidate keyword sentences, wherein each candidate keyword sentence has a mark, and the mark is associated with the category of the candidate keyword sentence; (b) according to the plurality of candidate keyword sentences and belongings The mark selects the corresponding search field of each candidate keyword sentence and the corresponding weighting parameter; (c) searches the plurality of legal data of the legal database according to the plurality of candidate keyword sentences to obtain the plurality of candidate legal materials, and obtains the plurality of candidate legal materials The legal data is used to calculate the relevance scores of each candidate legal data; and (d) the relevance ranking operation is performed according to the relevance scores of the plurality of candidate legal data, and the plurality of candidate legal data sorted by the relevance degree is provided.

In order to achieve the above objective, a more general implementation of the present case provides a method for searching for legal data for a legal data search system for performing a legal data search operation, wherein the legal data search system includes a search server and a search servo. The processor comprises a processor architecture for performing a law data search operation, and the legal data search processing method comprises the steps of: (a) receiving a keyword group, and expanding one of the keyword groups or the plurality of keyword sentences to obtain a plurality of candidates a keyword sentence; and (b) searching for a plurality of legal data of the legal database based on the plurality of candidate keyword sentences to obtain a plurality of candidate legal materials that are related to the keyword group.

In order to achieve the above object, a broader aspect of the present invention is to provide a computer readable memory medium, which is a storage program, and retrieves a server system after the retrieval server of the legal data search system loads and executes the program. Perform the legal data search processing method as described above.

1‧‧‧Law data search system interface

10‧‧‧ field

2‧‧‧Legal Information Search System

21‧‧‧Search server

200‧‧‧Host

210‧‧‧ processor

211‧‧‧ memory

2111‧‧‧Program

2112‧‧‧ Extended vocabulary

2113‧‧‧Category-Search Field-weight correspondence database

2114‧‧‧weight parameter database

212‧‧‧Communication unit

213‧‧‧Input and output unit

214‧‧‧Keyword tagging and selective extension module

216‧‧‧Keyword Search Field and Weighting Module

217‧‧‧Search and relevance score calculation module

218‧‧‧Sorting module

219‧‧‧ data output module

219a‧‧‧Highlighting unit

219b‧‧‧Content capture unit

22‧‧‧Law Database

221‧‧‧Legal information

222‧‧‧Information field

23‧‧‧User device

23a‧‧‧Smart Phone

23b‧‧‧ Tablet PC

23c‧‧‧Note Computer

23d‧‧‧Tablet computer

231‧‧‧Intelligent search interface

232‧‧‧Search field

233‧‧‧Search button

234‧‧‧Sort switch button

Steps of the search and processing method of the S10~S15‧‧  legal data search system

Steps for calculating the correlation score of S21~S23‧‧

Figure 1 is a schematic diagram of the interface used by the conventional legal data search system.

Figure 2A is a block diagram showing an exemplary embodiment of a legal data search system of the preferred embodiment of the present invention.

2B is a block diagram showing another exemplary embodiment of the legal data search system of the preferred embodiment of the present invention.

Figure 3 is a schematic diagram of the intelligent retrieval interface of the user device of Figure 2A.

Figure 4 is a schematic diagram of the architecture of the retrieval server of Figure 2A.

Figure 5 is a schematic diagram of the legal database of Figure 2A.

Figure 6 is a block diagram showing the architecture of the processor and memory of the retrieval server of Figure 4.

Figure 7 is a flow chart showing the search processing method of the legal data searching system of the preferred embodiment of the present invention.

Figure 8 is a schematic diagram showing the results of the retrieval of legal data by the user device of Figure 2A.

Figure 9 is a flow chart of a method for calculating the relevance score of the preferred embodiment of the present invention.

Some exemplary embodiments embodying the features and advantages of the present invention are described in detail in the following description. It is to be understood that the present invention is capable of various modifications in various aspects, and is not intended to limit the scope of the invention.

2A is a schematic structural diagram of an exemplary embodiment of a legal data search system of a preferred embodiment of the present invention, and FIG. 2B is a schematic structural diagram of another exemplary embodiment of a legal data search system of the preferred embodiment of the present invention, and FIG. 2A is a schematic diagram of the intelligent retrieval interface of the user device, FIG. 4 is a schematic diagram of the architecture of the retrieval server of FIG. 2A, and FIG. 5 is a schematic diagram of the legal database of FIG. 2A. As shown in Figures 2A, 2B, 3, 4 and 5, the legal data search system 2 of the present invention includes a retrieval server 21, a legal database 22, and at least one user device 23. In an embodiment, as shown in FIG. 2A, the search server 21 and the legal database 22 can be disposed in a host 200 and connected and communicated through a local device such as a connection port in the host 200. In another embodiment, as shown in FIG. 2B, the retrieval server 21 and the legal database 22 can be separate and separate physical devices and communicate through the communication network 20 Link and communication. The search server 21 and the at least one user device 23 are connected and communicated by the communication network 20, wherein the communication network 20 can be, for example, but not limited to, an internet, a mobile communication network, or the like. The user device 23 includes, but is not limited to, a smart phone 23a with a networking function, a tablet 23b, a notebook computer 23c, or a desktop computer 23d. The user device 23 is coupled to the retrieval server 21 via the communication network 20, and loads and displays an intelligent retrieval interface 231 for the user to input a keyword group and initiate a retrieval operation. As shown in FIG. 3, in the present embodiment, the intelligent search interface 231 includes a single search field 232 and a search button 233, wherein the search field 232 is used for the user to input the keyword group of the legal data to be searched. The search button 233 is for the user to click to initiate a search operation based on the keyword group entered in the search field 232. As shown in FIG. 5, the legal database 22 includes a plurality of legal materials 221, wherein each legal material 221 includes a plurality of data fields 222. As shown in FIG. 4, the retrieval server 21 includes a processor 210, a memory 211, a communication unit 212, and an input and output unit 213. The processor 210 is architected to perform a retrieval operation and to manage the retrieval server 21. The memory 211 is structured to store programs and store pre-stored data required to perform retrieval operations. The communication unit 212 is configured to be connected and communicated with the user device 23 and the legal database 22. The input and output unit 213 is structured to perform input and output of instructions and data.

FIG. 6 is a schematic diagram showing the architecture of the processor and the memory of the retrieval server of FIG. 4, and FIG. 7 is a flowchart of the search processing method of the legal data searching system of the preferred embodiment of the present invention. As shown in Figures 2A through 7, in the present embodiment, the processor 210 of the retrieval server 21 includes a keyword tag and selective expansion module 214, a keyword search field and weighting module 216, and a search and association. The score calculation module 217, the sorting module 218, and the data output module 219. The memory 211 of the retrieval server 21 includes a program 2111, an extended vocabulary 2112, a category-search field-weight correspondence database 2113, and a weight parameter database 2114. The legal data search system 2 of the present case and its search processing method are described below. First, as in step S10 As shown, when receiving the keyword group originating from the user device 23 and initiating the search job request, the processor 210 of the search server 21 analyzes the keyword group and starts executing the legal data search operation. For example, if the input keyword group is "Taiwan No. 1442 Tainan Court," the processor 210 of the retrieval server 21 parses and extracts the keyword group into a plurality of keys. Words such as "Letter No. 1442" and "Tainan District Court".

Then, as shown in step S11, the processor 210 executes the keyword tagging and selective expansion module 214 to mark each keyword sentence of the keyword group, so that each keyword sentence is marked. Candidate keyword sentence. In this step, the marking job may include, for example, but not limited to, marking the category of each candidate keyword sentence belonging to the court name, the judgment font size or the regulation, in other words, the marking of each candidate keyword sentence and the candidate The category of the keyword sentence is related. In this step, the processor 210 may further perform an extended operation on each keyword sentence of the keyword group, wherein the selective means that the processor 210 expands or does not expand each keyword sentence of the keyword group. operation. The foregoing expansion operations may include, for example, but not limited to, synonym expansion, digital to national expansion, national to digital expansion, regulatory name extension, regulatory article extension, case extension, case identification and extension, and court identification and extension. Thereby, the expanded plurality of keyword sentences are obtained. In some embodiments, the processor 210 may first perform an extended job to expand the keyword sentence of the keyword group to obtain the expanded plurality of keyword sentences, and then the processor 210 performs the marking operation and is expanded. The latter plurality of keyword sentences become a plurality of candidate keyword sentences each with a mark. For example, one of the above keyword groups, the wording "Symbol No. 1442", was expanded into "Keywords No. 1442" and "Voice No. 1442" and the like. Marked as "judgment font size" and become a candidate keyword sentence with a mark, and expand the other keyword sentence "Tainan Court" of the above keyword group into "Tainan District Court", "Tainan District Court", "Tainan" Key words such as "court" and "Tainan District Court" and mark it as "Law The name of the hospital becomes a candidate keyword sentence with a mark. In other embodiments, the processor 210 may perform a marking operation to mark each keyword sentence of the keyword group respectively, and then the processor 210 performs an extended job to expand the keyword sentence with the tag. In order to obtain a plurality of candidate keyword sentences with tags.

In some embodiments, the extended keyword library of the memory 211 may pre-store the extended keyword sentence of the extended job. Therefore, the keyword tag and the selective expansion module 214 may be extended words in the memory 211 when performing the extended job. The corresponding extended keyword sentence is obtained in the library 2112. It should be emphasized that the marking and selective expansion operations of this step are not limited to the above embodiments, and may be changed according to actual application requirements, thereby increasing the breadth of the keyword search range and achieving association. The effect of sexual search. In this embodiment, the categories of legal materials may include, for example, but are not limited to, court names, criminal judgments, civil judgments, administrative judgments, judgment font numbers, judgment cases, judgment dates, referee texts, regulatory names, or regulatory texts. Since the markup of each candidate keyword sentence is related to the category of the candidate keyword sentence, the category of the candidate keyword sentence can be determined according to the mark of the candidate keyword sentence.

Then, as shown in step S12, after completing the marking operation, the processor 210 executes the keyword search field and the weighting module 216 to calculate each of the plurality of candidate keyword sentences and their respective categories. A candidate keyword sentence performs search field and weight adjustment and assignment. For example, the above-mentioned candidate keyword sentence "Voice No. 1442" according to its mark can be known as the "judgment font size", and after performing the search field and the weight adjustment and assignment operation, the candidate keyword can be used according to the candidate keyword. The sentence and its category, the data field of the set of candidate keyword sentences to be searched in the legal database 22 may include, for example, but not limited to, the "judgment font size" of the legal data 221 and the data field of the "judgment main text". (See Figure 5), which expands the search breadth of keywords. In addition, a corresponding weighting parameter is given depending on the importance of the category and/or field. In some embodiments, the memory 211 The category-search field-weight corresponding database 2113 can pre-store the category-search field-weight corresponding data, so the keyword search field and weighting module 216 performs the search field and weight adjustment and allocation. At the time of the homework, the category-search field-weight corresponding data can be obtained in the category-search field-weight corresponding database 2113 of the memory 211. It should be emphasized that the search field and the weight adjustment and allocation operation in this step are not limited to the above embodiments, and may be changed according to actual application requirements, thereby increasing the breadth of the keyword search range, and Achieve the ability to search by relevance.

Then, as shown in step S13, after completing the search field and the weight adjustment and assignment operation, the processor 210 executes the search and relevance score calculation module 217 to select each of the candidate keyword sentences according to the selected one. Or a plurality of search columns are located in the plurality of legal materials 221 of the legal database 22 for searching to filter out the plurality of candidate legal materials 221 . At the same time, the relevance score calculation is performed according to the given weighting parameter and a specific algorithm. For example, for the legal data 211 obtained by any of the search filters, the candidate keyword sentence "Voice No. 1442" appears in the "judgment font number" field of the legal data 211 according to the field. The weight value obtained by calculating the weighting parameter wa1 corresponding to the given weighting parameter wa1 is a1, and the weight value obtained by the weighting parameter wa2 corresponding to the field corresponding to the field in the "judgment main text" field of the grammar data 211 is a2, so In the case of the legal data 211, the total weight value of the candidate keyword sentence "V. No. 1442" is a1+a2. Similarly, for the grammar data 211, the other candidate keyword sentences described above also calculate their total weight values in the same manner. Then, the total weight values obtained by calculating all the candidate keyword sentences are added together to obtain the relevance score of the legal data 211. The plurality of candidate candidate legal materials 211 obtained by the search and filtering are respectively calculated and obtained by the above method. In this embodiment, the relevance score of each candidate legal data 211 is the basis for determining the relevance of the candidate legal data and the keyword group, and the higher the relevance score represents the candidate legal data and the keyword. The higher the association degree of the group; on the contrary, the lower the relevance score The lower the degree of association between the candidate legal data and the keyword group. It should be emphasized that the search and relevance score calculation operation of this step is not limited to the above embodiment, and may be changed according to actual application requirements.

In some embodiments, if the candidate legal data 211 is a special, high-value legal material such as a "criteria" or a "high-value referee", the relevance score of the candidate legal data is in accordance with the legal data. The nature of its own further adds its weight scores to increase the relevance of the candidate legal data with particularity and high value. The aforementioned "jurisprudence" refers to a judgment that has already taken effect. The court used the precedent as a precedent in the judgment of similar cases. Such precedents for such judgments are called "jurisprudence". It should be emphasized that the weighting of the legal data addition weights in this step is not limited to "jurisprudence" or "high-value referees". Other legal materials with special and high value may also be included in the reference.

Thereafter, as shown in step S14, after completing the search and relevance score calculation operation, the ranking module 218 of the retrieval server 21 is executed to perform the sorting operation according to the relevance score of the plurality of candidate legal materials 211, that is, according to The relevance scores of the plurality of candidate legal data 211 are sorted from high to low in order to rank the candidate legal data 211 with the highest relevance score (ie, the highest relevance) in the first stroke, and the relevance score is second highest (ie, association) The candidate legal data of the higher degree is ranked second, and so on, to generate one of the relevance rankings to retrieve the query results.

Finally, as shown in step S15, after completing the relevance ranking job, the data output module 219 of the server 21 is executed to output the candidate legal data 211 sorted according to the relevance degree, and the communication network 20 The search results are provided and displayed on the user device 23 for viewing by the user, as shown in FIG. In some embodiments, the data output module 219 of the processor 211 further includes a highlight indicator unit 219a to highlight the position of the candidate keyword sentence in each candidate legal material 211 (for example, but not limited to gray Presentation of the order, color, bold or bottom line) for the user to quickly view News. In other embodiments, the data output module 219 of the processor 211 further includes a content capturing unit 219b for presenting relatively important paragraphs in each candidate legal material 211 (eg, the candidate legal data has A field or paragraph with a higher weight value, so that the user can quickly view the desired information.

In some embodiments, the smart search interface 11 of the user device 23 may further include a sort switch button 234 (as shown in FIG. 8) when the page for displaying the query result is displayed, wherein the sort switch button 234 is available for use. Click to switch to sort the results based on the date, so that users can flexibly apply. In the embodiment, when the processor 211 receives the sorting switching request from the user device 23, the sorting module 218 of the processor 211 is executed to use the plurality of candidate legal materials 211 according to the date (eg, the date of the judgment, The sorting operation is performed on the publication date and the announcement date. Thereafter, the time-ordered search results are provided to the user device 23 for display by the user through the data output module 219. In some embodiments, the intelligent search interface 11 of the user device 23 may further include a recommended keyword group link (not shown) when displaying the page of the query result to simultaneously display the keyword groups that may be related. At least one search link. When the result of the query is not the legal information that the user wants to search, the user can click the corresponding search link of the recommended keyword group provided by the system to perform other search operations of the keyword group that may be associated, and the convenience of use is achieved. Sex. In this embodiment, the keyword tag and selective extension module 214 of the processor 211 can provide the candidate keyword group to the recommended keyword group link of the intelligent search interface 11 of the user device 23 in a connected manner.

According to the concept of the present case, the calculation method of the relevance score is not limited to the above embodiment. Figure 9 is a flow chart for calculating the relevance score of the preferred embodiment of the present invention. As shown in Fig. 9, the method for calculating the relevance score of the present case includes the following steps. First, as shown in step S21, the candidate keyword sentence is judged to be a referee class or a statute class based on the flag attached to the candidate keyword sentence. For example, the candidate keyword sentence "Corruption Crimes Ordinance" carries the mark on behalf of "Law According to the mark, it can be judged that the candidate keyword sentence belongs to the category of "regulatory name" and further classified as a statute; relatively, the candidate keyword sentence "with the Corruption Ordinance" The mark system represents the "judgment case", so according to the mark, the candidate keyword sentence can be judged to belong to the "judgment case" and further classified as a referee class. The above description is made by way of example, and the manner of judging the category is not limited thereto, and may be changed according to actual conditions. Then, as shown in step S22, the search and relevance score calculation module 217 of the processor 211 provides respective specific relevance score algorithms corresponding to the referee class or the rule class according to the foregoing judgment result, in other words, if the candidate keyword The sentence is classified as a statute, and the legal data is calculated by a law of relevance calculation, and the relevance score of the candidate keyword in the legal data is obtained; if the candidate keyword is classified as a referee For the class, the legal data is calculated by a referee correlation calculation program, and the relevance score of the candidate keyword sentence in the legal data is obtained. Finally, as shown in step S23, all the candidate keyword sentences are summed over all the relevance scores of the candidate legal data, and the relevance score of the candidate legal data is obtained.

Please continue to refer to Figure 9, as stated above, the rule correlation score calculation program is operated by the following formula:

Among them, Max ( matchWeight ) is the largest score, which is applicable in the case where the name is completely consistent; matchWeight is the score in the case of generally conforming to the rule (meaning that it is not the maximum weight); tieBreak is a weighted parameter; n is the total number of search conditions, and the correlation score (Score) of the legal data is calculated by the operation of this formula. For example, please refer to Table 1. If the keyword sentence entered is "Criminal Law" and the legal data is "Criminal Law of the Republic of China", the input keyword sentence "Criminal Law" is marked and expanded to become "Republic of China." Candidate keyword sentences such as "criminal law" and "criminal law", when the search condition is "criminal law" or "Criminal Law of the Republic of China", the legal data "Criminal Law of the Republic of China" meets the search conditions, and the name is completely consistent, so the maximum score Max is applied. ( matchWeight ). Furthermore, when the search conditions are "the beginning of "criminal", and the end of "law" or "regulation", "including "penalty", "law" and appear in order", etc., the legal data The "Criminal Law of the Republic of China" is also in line with the search conditions, and the scores in the general compliance with the rules are applied, and the calculations are summed according to the formula 1. The correlation scores of the legal data "the Republic of China Criminal Law" are obtained (see Table 1). Shown).

Please continue to refer to Figure 9, as stated above, the referee relevance score calculation program is operated by the following formula:

Where queryNorm _i,j is a normalized search score, and its calculation formula is:

Tf _i,j is the number of times the candidate keyword sentence i appears in the field j, and its calculation formula is:

Idf _i,j is the rarity of the candidate keyword sentence i in all legal materials, and its calculation formula is: (Formula 5) idf _i,j =1+ log (total number of files / (including total number of keywords + 1))

The fieldweight is the weighting parameter of the search field j to distinguish the importance of the field. Through the calculation of this formula, the relevance score (Score) of the legal data is calculated. For example, please refer to Table 2. If the keyword entered is "Civil Service Uses Corruption" and the legal information is "Judgement No. 839 of the 99th Annual Appeal of the Taiwan High Court Tainan Branch", then its The correlation score calculation is explained as follows. The candidate keyword sentences that are marked and expanded may include, for example, but not limited to, candidate keyword sentences such as "civil servant use position" and "corruption", and the search field is searched for "case" and "referee text". Among them, the candidate keyword sentence "Civil Service Use Title" does not appear in the search field "Case", so the queryNorm, tf and idf values are all 0; the candidate keyword sentence "Corruption" appears once in the search field "Case" Calculate tf =1 through Equation 4, and then calculate queryNorm =0.94229126 via Equation 3, and calculate the rarity idf =6.9492507 of "corruption" in all legal materials by Equation 5. Then, calculate the queryNorm, tf, and idf values of the candidate keyword sentence "Civil Service Use Title" in the search field "referee text" and the candidate keyword sentence "corruption" in the search field "referee text". The method is similar to the calculation method described in the first half of this paragraph, so it will not be repeated here. Finally, after all the calculations of queryNorm, tf and idf are completed, the correlation scores of the 99th Annual Appeal No. 839 of the Taiwan High Court Tainan Branch (Taiwan High Court Tainan Branch Criminal Judgment No. 839) can be obtained by the calculation of Equation 2 (the judgment). Score) = 120.235306.

In summary, the legal data search system and search processing method of the present case can provide relevant legal data related to the keyword, improve the accuracy of the search, provide a friendly and easy-to-operate user experience, and save the use. The time of retrieval and reading.

This case has been modified by people who are familiar with the technology, but it is not intended to be protected by the scope of the patent application.

Steps of the search and processing method of the S10~S15‧‧  legal data search system

Claims (23)

A legal data search system comprising: a legal database for storing a plurality of legal data; a client device comprising an intelligent search interface, wherein the intelligent search interface is for a user to input a keyword group; and a search servo The device is connected to and communicates with the legal database and the client device, and includes a processor, wherein the processor performs a legal data searching operation according to the keyword group of the client device, and the processor includes: a keyword tag and a selective extension module, performing one tag operation on the one or more keyword sentences of the keyword group and selectively performing an extended job to obtain a plurality of candidate keyword sentences, wherein each candidate The keyword sentences respectively have a mark, and the mark is associated with one of the candidate keyword sentences; a keyword search field and a weight allocation module are selected according to the plurality of candidate keyword sentences and the tag to which the tag belongs Corresponding search fields and corresponding weighting parameters of each candidate keyword sentence; a search and relevance score calculation module, according to the complex The candidate keyword sentences search the plurality of legal materials of the legal database to obtain a plurality of candidate legal materials, and calculate the relevance scores of the candidate legal materials for the plurality of legal data, wherein the relevance scores The sum of the weight values of the plurality of candidate keyword sentences, wherein the weight value of each candidate keyword sentence is calculated by at least one parameter, and the at least one parameter includes a corresponding search appearing in the candidate legal data. a parameter of the number of times of the field, a rarity parameter of the candidate keyword sentence, and a weighting parameter of the corresponding search field of the candidate keyword sentence; and a sorting module, according to the relevance score of the plurality of candidate legal data a relevance ranking job, providing a plurality of candidate legal data sorted by relevance to the User device. The legal data search system of claim 1, wherein the search server and the legal database are co-located on a host and connected and communicated through at least one connection, wherein the search server and the search server The client devices are respectively connected to a communication network, and the retrieval server and the client device are connected and communicated through the communication network. The legal data searching system of claim 1, wherein the search server, the legal database, and the client device are respectively connected to a communication network, and the search server, the legal database, and the user The end device is connected and communicated through the communication network. The legal data search system of claim 1, wherein the intelligent search interface comprises a single search field and a search button, wherein the search field is structured by inputting the keyword group, and the search key system is Start the legal data search operation. The legal data searching system of claim 1, wherein the search server further comprises a memory, the memory comprising a program, an extended dictionary, a category-search field-weight correspondence database, and a weight Parameter database. The legal data searching system of claim 1, wherein the processor further comprises a data output module, wherein the plurality of candidate legal data sorted by the relevance degree is output, and the communication data is used by a communication network. The legal data retrieval result is provided to the user device display. For example, the legal data search system described in claim 1 of the patent scope includes the expansion of synonymous words, the expansion of digital to national characters, the expansion of national characters to digital, the expansion of legal names, the extension of laws and regulations, and the expansion of the nature of cases. Case identification and extension or court identification and extension. For example, the legal data search system mentioned in the first paragraph of the patent application, wherein the category is the court name, criminal judgment, civil judgment, administrative judgment, judgment font number, judgment case, judgment date, referee main text, legal name or regulatory text. . The legal data searching system of claim 1, wherein the sorting module sorts the relevance scores of the plurality of candidate legal materials in descending order. The legal data searching system of claim 1, wherein the data output module of the processor includes a highlighting unit to display the plurality of candidate keyword sentences in the plurality of candidate legal materials. The location is marked. The legal data search system of claim 1, wherein the data output module of the processor comprises a content extraction unit for performing part of the candidate legal materials of the plurality of candidate legal materials. The capture is displayed on the user device. A legal data search processing method for a legal data search system for performing a legal data search operation, wherein the legal data search system includes a search server, the search server including a processor architecture for performing the legal data search operation The legal data search processing method includes the steps of: (a) receiving a keyword group, performing a marking operation on one or a plurality of keyword phrases of the keyword group, and selectively performing an extended job to obtain a plurality of candidates a keyword sentence, wherein each of the candidate keyword sentences respectively has a mark, and the mark is associated with one of the candidate keyword sentences; (b) is selected according to the plurality of candidate keyword sentences and the associated mark Corresponding search fields and corresponding weighting parameters of each candidate keyword sentence; (c) searching for a plurality of legal data of a legal database according to the plurality of candidate keyword sentences to obtain a plurality of candidate legal materials, and the plurality of candidate legal materials The brush law data is used to calculate the relevance scores of each candidate legal data; and (d) based on the plurality of candidate legal data The relevance score is performed by a relevance ranking operation, wherein the relevance degree score of the candidate legal data is a sum of weight values of the plurality of candidate keyword sentences, and the weight value of each candidate keyword sentence is determined by at least one parameter. Calculating, the at least one parameter includes a parameter parameter of a corresponding search field appearing in the candidate legal data, a rarity parameter of the candidate keyword sentence, and a weighting of the corresponding search field of the candidate keyword sentence Parameters, 俾 provide multiple candidate legal data sorted by relevance. The method for searching for a legal data as described in claim 12, wherein the processor includes a keyword tag and a selective expansion module, a keyword search field and a weighting module, a search and a relevance score. a computing module and a sorting module, wherein the step (a) is performed by the keyword tag and the selective expansion module, and the step (b) is performed by the keyword search field and the weighting module, Step (c) is performed by the search and relevance score calculation module, and the step (d) is performed by the sorting module. The method of searching for a legal data as described in claim 12, wherein in the step (a), the marking operation is performed prior to performing the expanding operation. The method for searching for legal data according to claim 12, wherein in the step (a), the expanding operation is performed prior to performing the marking operation. The method for searching for a legal data as described in claim 12, further comprising the step (e): outputting the plurality of candidate legal data sorted by the relevance degree, and searching a legal data through a communication network. The result is provided to a user device display. The method for searching for a legal data as described in claim 16 , wherein the step (e) further comprises the step of: (e1) marking the position of the plurality of candidate keyword sentences in the plurality of candidate legal materials; And (e2) extracting a portion of the legal data of the plurality of candidate legal materials for display on the user device. For example, the legal data searching and processing method described in claim 12, wherein the expansion operation includes synonym expansion, digital to national expansion, national character to digital expansion, regulatory name extension, regulatory article extension, and case extension , identification and extension of case numbers or court identification and extension. For example, the legal data search and processing method described in claim 12, wherein the category is a court name, a criminal judgment, a civil judgment, an administrative judgment, a judgment font number, a judgment case, a judgment date, a referee main body, a law name, or a regulation. Text. The method for searching for a legal data according to claim 12, wherein in the step (d), the ranking operation is sequentially sorted according to the degree of relevance of the plurality of candidate legal data from high to low. A legal data search processing method for a legal data search system for performing a legal data search operation, wherein the legal data search system includes a search server, the search server including a processor architecture for performing the legal data search operation The legal data search processing method includes the steps of: (a) receiving a keyword group, performing an extended operation on one or a plurality of keyword sentences of the keyword group to obtain a plurality of candidate keyword sentences; (b) The plurality of candidate keyword sentences search for a plurality of legal data of a legal database to obtain a plurality of candidate legal materials that are related to the keyword group, and perform the candidate legal data for the plurality of legal materials. Correlating degree score calculation; and (c) performing a relevance ranking operation according to the relevance score of the plurality of candidate legal data, wherein the relevance degree score of the candidate legal data is the sum of the weight values of the plurality of candidate keyword sentences The weight value of each candidate keyword sentence is calculated by at least one parameter, and the at least one parameter includes The number parameter of the corresponding search field of the candidate legal data, the rarity parameter of the candidate keyword sentence, and the weighting parameter of the corresponding search field of the candidate keyword sentence, and the plurality of times sorted by the relevance degree Candidate legal data. For example, the legal data search processing method described in claim 21, wherein the expansion operation includes synonym expansion, digital to national expansion, national character to digital expansion, regulatory name expansion, regulatory article extension, and case extension , identification and extension of case numbers or court identification and extension. A computer readable memory medium for storing a program, and after the retrieval server loads and executes the program in one of the legal data search systems, the retrieval server performs the method as described in claim 12 Legal data search processing method.