Intellectual property information display method and system

Abstract

The invention discloses an intellectual property information display method and system, relates to the technical field of data processing, and solves the problems that a scientific and reasonable screening and classification method is lacked, retrieval results cannot be effectively screened and classified according to key information in retrieval conditions, a user is difficult to focus key contents quickly, and the user experience is improved. According to the method, a trained MemoRAG large language model is used for deeply splitting retrieval conditions, keywords are accurately marked by means of a professional vocabulary, a system can accurately understand the retrieval intention of a user, the accuracy and correlation of retrieval results are improved, classification is carried out by calculating the proportion of the number of the keywords in the screening results, and the retrieval efficiency is improved. According to the method, a user is helped to view and analyze intellectual property information from different dimensions, meanwhile, different signals are generated according to the proportion of the number of browsing operations in combination with historical data of the user, different display modes are guided respectively, and personalized display of retrieval results is achieved.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a method and system for displaying intellectual property information. Background Art

[0002] In the field of intellectual property retrieval, as the number of intellectual property rights continues to grow, users' requirements for retrieval accuracy and efficiency are increasing.

[0003] According to a patent application with publication number CN115795024A, a method and system for displaying intellectual property information are disclosed, the method comprising: obtaining search conditions and user information, the user information at least comprising historical search records; performing an intellectual property search according to the search conditions to obtain intellectual property data; classifying the historical search records according to the intellectual property data to obtain multiple groups of historical search data; performing feature extraction on the intellectual property data, determining an intellectual property information display scheme, and displaying the intellectual property data.

[0004] Traditional search methods have many shortcomings when dealing with complex search conditions, such as the inability to accurately understand the search intent, resulting in inaccurate search results, and a large amount of irrelevant information interfering with users' acquisition of valid content. At the same time, existing technologies also have defects in the screening, deduplication and personalized display of search results, making it difficult to meet users' needs for in-depth analysis of intellectual property information from different dimensions, and failing to fully combine users' historical behavior data to realize personalized search services. Summary of the invention

[0005] In view of the deficiencies in the prior art, the present invention provides a method and system for displaying intellectual property information, which solves the problems of lack of scientific and reasonable screening and classification methods, inability to effectively screen and classify search results according to key information in the search conditions, and difficulty for users to quickly focus on key content, thus affecting search efficiency.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: a method for displaying intellectual property information, the method specifically comprising the following steps: Obtain results and corresponding conditions based on search conditions, split search conditions and mark keywords using a large language model, filter and classify search results, and obtain classified filtering results; Select any set of classification screening results to determine duplicate data, retain the latest one as the standard result, hide the rest, and process all groups to obtain the final standard result set; Extract keywords from the screening results without duplicate data, calculate the key percentage value, sort by this value to generate sorting information, and do this for all classification screening results; Obtain sorting information and user history data, divide into segments by time T, count the percentage of user browsing operations in each segment, and compare with the preset value; If the proportion is greater than the preset value, a quantitative display analysis signal is generated and the results are displayed in combination with relevant information; if it is less than the preset value, a comprehensive display analysis signal is generated and the results are displayed as required.

[0007] As a further solution of the present invention, the specific method of obtaining the classification screening results is: Obtain the corresponding search results according to the search conditions, and at the same time obtain the search conditions corresponding to the search results, use the large language model to split the search conditions to obtain split information, and mark the keywords, and at the same time filter the search results according to the split information to generate filtered results, and at the same time classify the filtered results based on the proportion of the number of keywords in the search conditions to generate classified filtering results.

[0008] As a further solution of the present invention, the specific method of obtaining the final standard result set is: Select any group of classification and screening results as the analysis object, and obtain the search results therein, determine the duplicate data based on the similarity of the search result content, retain the latest one in the duplicate data as the standard result, and mark the rest as hidden results. All classification and screening results are processed in this way to obtain the final standard result set.

[0009] As a further solution of the present invention, the specific method of generating the sorting information is: Get the filtering results that do not contain duplicate data, extract their keywords, calculate the ratio of the number of keywords in each filtering result to the total number of keywords, sum these ratios to get the key ratio value, sort the filtering results from large to small according to the key ratio value, generate sorting information, repeat this operation for all classified filtering results, and generate their respective sorting information.

[0010] As a further solution of the present invention, the specific method of counting the proportion of the number of user browsing operations in each segment and comparing it with the preset value is: Obtain sorting information and retrieve user historical data, segment the historical data with time T as the period, obtain b historical period segments a, count the user browsing operations in each period, find out the operations with the same number of views, count the corresponding number of search results, calculate their proportion, and compare it with the preset number proportion set by the operator.

[0011] As a further solution of the present invention, the specific method of generating a quantity display analysis signal and displaying the result in combination with relevant information is: If the proportion is greater than the preset value, a quantity display analysis signal is generated, display information is generated according to the same number of browsing operations, the cosine similarity between the screening results and the search conditions is calculated and sorted, and the results are displayed in combination with the display information.

[0012] As a further solution of the present invention, the specific method of generating a comprehensive display analysis signal and displaying the result as required is: If the proportion is less than the preset value, a comprehensive display analysis signal is generated, the browsing operation with the largest proportion of quantity is found, and the results are sorted by cosine similarity between the screening results and the search conditions, and displayed.

[0013] An intellectual property information display system, comprising: The information acquisition module is used to obtain the corresponding search results according to the search conditions, and at the same time obtain the search conditions corresponding to the search results, and transmit them to the search analysis module; The retrieval analysis module is used to use the large language model to split the retrieval conditions to obtain split information and mark keywords, and at the same time, filter the retrieval results according to the split information to generate filtering results, and classify the filtering results according to the proportion of the number of keywords in the retrieval conditions to generate classified filtering results, and select a group of classified filtering results as the analysis object, and obtain the retrieval results therein, determine the duplicate data according to the similarity of the retrieval results content, retain the latest time in the duplicate data as the standard result, and mark the rest as hidden results, and process all the classified filtering results in this way to obtain the final standard result set, and transmit the final standard result set to the secondary analysis module; The secondary analysis module is used to obtain the screening results that do not contain duplicate data, extract its keywords, calculate the ratio of the number of keywords in each screening result to the total number of keywords, sum these ratios to obtain the key ratio value, sort the screening results from large to small according to the key ratio value, generate sorting information, repeat this operation for all classified screening results, and generate their respective sorting information; Obtain sorting information and retrieve user historical data, segment the historical data with time T as the period, obtain b historical period segments a, count the user browsing operations in each period, find out the operations with the same number of views, count the number of corresponding search results, calculate their proportion, and compare with the preset number proportion set by the operator; If the proportion is greater than the preset value, a quantity display analysis signal is generated, display information is generated according to the number of the same browsing operations, the cosine similarity between the screening results and the search conditions is calculated and sorted, and the results are displayed in combination with the display information; If the proportion is less than the preset value, a comprehensive display analysis signal is generated, the browsing operation with the largest proportion of quantity is found, the screening results are sorted by cosine similarity with the search conditions, the results are displayed, and the display results are transmitted to the display output module at the same time; The display output module is used to display the display results to the corresponding search users.

[0014] The present invention provides a method and system for displaying intellectual property information. Compared with the prior art, it has the following beneficial effects: The present invention uses the trained MemoRAG large language model to deeply split the search conditions, and uses the professional vocabulary to accurately mark keywords, so that the system can accurately understand the user's search intention, improve the accuracy and relevance of the search results, and classify them by calculating the proportion of the number of keywords in the screening results. It helps users view and analyze intellectual property information from different dimensions, quickly focus on key content, and improve search efficiency and user experience.

[0015] Based on the text similarity algorithm, duplicate data is judged, and the ones with the highest time priority are retained as standard results, and the rest are marked as hidden results, reducing redundant information, improving the quality and availability of search results, calculating the key proportion value to sort the screening results, and generating comprehensive and orderly sorting information. At the same time, combined with user historical data, different signals are generated according to the proportion of browsing operations, guiding different display methods respectively, realizing personalized display of search results, giving priority to displaying results that meet user browsing preferences and have high similarity with search conditions, and providing users with better search services. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a step method diagram of the present invention; Figure 2 This is a block diagram of the system principle of the present invention. DETAILED DESCRIPTION

[0017] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0018] For example, see Figure 1 The present application provides a method for displaying intellectual property information, which specifically includes the following steps: Step S1: Based on the search conditions input by the user, the system starts the search program and obtains the corresponding search results from the huge intellectual property database. While obtaining the search results, the system will record the search conditions corresponding to each search result. The search conditions here not only include the original information input by the user, but also specifically relate to the display position information of the search results in the display interface, providing a basis for subsequent sorting and display.

[0019] Step S2: After the system obtains the search conditions input by the user, it calls the pre-trained MemoRAG large language model to perform in-depth segmentation. With its powerful language understanding ability, MemoRAG can decompose complex search conditions into multiple segmentation information with clear semantics. Subsequently, the system uses a pre-built professional vocabulary to accurately mark the keywords in the segmentation information. The vocabulary covers a wealth of intellectual property terms, various common technical terms, and widely recognized conventional expressions in the industry to ensure the accuracy and comprehensiveness of keyword marking.

[0020] The marked keywords are carefully matched and screened with the search results. The system checks each search result one by one. If the search result contains at least one marked keyword, it will be retained; if the search result does not contain any marked keywords, it will be removed. After this screening process, a set of screening results that meet the preliminary requirements are obtained and numbered in sequence, recorded as i (i = 1, 2, …, j), where j represents the total number of screening results.

[0021] For each screening result i, the system counts the number of keywords contained in it and calculates the ratio of this number to the total number of all tagged keywords, that is, the quantity ratio. For example, if there are 5 tagged keywords in total, and a screening result contains 3 keywords, its quantity ratio is 3÷5 = 0.6. The system classifies the screening results according to the quantity ratio, classifies the screening results with the same quantity ratio into the same category, and generates classified screening results. This classification method helps users view and analyze intellectual property information from different dimensions and quickly focus on key content.

[0022] Step S3: Conduct an in-depth analysis of the generated classification and screening results, focusing on the duplicate data. The system randomly selects a group of classification and screening results as the analysis object, comprehensively obtains all search results in the group, and uses advanced text similarity algorithms to determine whether there are duplicate data based on the content similarity of the search results. The operator can set a preset value of content similarity according to actual needs, such as 80%. If the content similarity of two search results exceeds the preset value, they are determined to be duplicate data.

[0023] After determining the duplicate data, the system obtains the time information corresponding to these duplicate data. For duplicate data, the system retains the one with the highest time priority as the standard result, and marks the remaining duplicate data as hidden results. For example, in a set of classification and screening results, there are 10 search results. The content similarity of result 3 and result 7 exceeds 80% after algorithm analysis, which are duplicate data. If the release time of result 3 is January 1, 2022, and the release time of result 7 is May 1, 2021, since January 1, 2022 is later than May 1, 2021, the system will retain result 3 as the standard result and mark result 7 as a hidden result. Similarly, all classification and screening results are processed in the same way to obtain the corresponding standard result set.

[0024] The system obtains the screening results that do not contain duplicate data, and extracts the keywords corresponding to each screening result. The correlation value is obtained by calculating the ratio of the number of corresponding keywords in each screening result to the number of all keywords. These correlation values ​​are summed up to obtain the key ratio value. The system sorts the screening results from large to small according to the key ratio value and generates sorting information. This operation is performed on all classified screening results to generate comprehensive and orderly sorting information, which provides strong support for the subsequent accurate display of search results.

[0025] For example, if a total of 10 keywords are determined in this search, and a certain screening result contains 4 of them, then the correlation value of the screening result is 4÷10=0.4. The correlation value of each screening result is summed up to obtain the key proportion value. Assuming that there are 3 screening results in the analysis object, and their correlation values ​​are 0.4, 0.3, and 0.2 respectively, then the key proportion value is 0.4+0.3+0.2=0.9, and the screening results are sorted from large to small according to the key proportion value. In the above example, after the screening results are sorted according to the key proportion value, the screening result with a key proportion value of 0.4 ranks first, followed by 0.3, and finally 0.2. Sorting information is generated based on the sorting results. The information can be presented in a list format, including the number of the screening results, the key proportion value, and other content.

[0026] Step S4: First, obtain the ranking information generated for the search results, which reflects the degree of correlation between the screening results and the search conditions. At the same time, collect the historical data of the search user, which covers important information such as the user's previous search records and browsing behaviors. Then, segment the historical data with time T as the cycle, and divide the historical data into multiple historical period segments, which are denoted as a in sequence. The value range of a is from 1 to b, and b represents the total number of historical period segments. For example, if the time period T is set to one week and the user's historical data for the past three months is analyzed, then b is approximately equal to 12 (assuming four weeks per month).

[0027] For each historical period, obtain the browsing operations of the searching user in the period in detail. The browsing operation here mainly focuses on the number of search results viewed by the user at one time. Assign a label k to each browsing operation, and k=1, 2, ..., h, where h represents the total number of browsing operations in the historical period. For example, in a certain historical period, the user viewed the search results 8 times in total, then h=8. Count the number of search results viewed in each browsing operation, and for browsing operations with the same number of views, summarize the corresponding number of search results. Then, calculate the proportion of these search results with the same number of views in the total number of search results viewed by all browsing operations in the historical period.

[0028] Compare the calculated quantity ratio with the preset quantity ratio value. The preset quantity ratio value is a reference standard set by the operator based on business needs, user behavior analysis and other factors. If the quantity ratio is greater than the preset quantity ratio, the system generates a quantity display analysis signal; if the quantity ratio is less than the preset quantity ratio, a comprehensive display analysis signal is generated. These two signals will guide different subsequent analysis and processing paths.

[0029] When a quantity display analysis signal is received, the signal is further analyzed to find out the specific circumstances corresponding to the same browsing operation, and the number of search results corresponding to these operations is obtained. The quantity display information is generated based on the quantity, which is used to intuitively present the user's attention to a specific number of search results. Then, all the screening results are processed, and the cosine similarity algorithm is used to calculate the similarity value between each screening result and the search condition. The screening results are sorted in order from large to small according to the similarity value, and the display results are generated in combination with the quantity display information. Finally, these display results are displayed to the corresponding search users in the order in which they are generated, so that users can give priority to results that are highly similar to the search conditions and meet their browsing quantity preferences.

[0030] If a comprehensive display analysis signal is received, the browsing operation with the largest number is first found and recorded as the operation to be analyzed. Taking the operation to be analyzed as the standard, the screening results are also sorted from large to small according to the similarity value between the screening results and the search conditions to generate the display results. Finally, the display results are displayed to the corresponding search users to help users quickly obtain the search results that comprehensively consider the historical browsing behavior and the matching degree of the search conditions.

[0031] For example 2, please refer to Figure 2 The present application provides an intellectual property information display system, including an information acquisition module, a search and analysis module, a secondary analysis module and a display output module, and combines Figure 2 It can be known that the above functional modules are electrically connected in a unidirectional manner.

[0032] The information acquisition module is used to obtain the corresponding search results according to the search conditions, and at the same time obtain the search conditions corresponding to the search results, and transmit them to the search analysis module; The retrieval analysis module is used to use the large language model to split the retrieval conditions to obtain split information and mark keywords, and at the same time, filter the retrieval results according to the split information to generate filtering results, and classify the filtering results according to the proportion of the number of keywords in the retrieval conditions to generate classified filtering results, and select a group of classified filtering results as the analysis object, and obtain the retrieval results therein, determine the duplicate data according to the similarity of the retrieval results content, retain the latest time in the duplicate data as the standard result, and mark the rest as hidden results, and process all the classified filtering results in this way to obtain the final standard result set, and transmit the final standard result set to the secondary analysis module; The secondary analysis module is used to obtain the screening results that do not contain duplicate data, extract its keywords, calculate the ratio of the number of keywords in each screening result to the total number of keywords, sum these ratios to obtain the key ratio value, sort the screening results from large to small according to the key ratio value, generate sorting information, repeat this operation for all classified screening results, and generate their respective sorting information; Obtain sorting information and retrieve user historical data, segment the historical data with time T as the period, obtain b historical period segments a, count the user browsing operations in each period, find out the operations with the same number of views, count the number of corresponding search results, calculate their proportion, and compare with the preset number proportion set by the operator; If the proportion is greater than the preset value, a quantity display analysis signal is generated, display information is generated according to the number of the same browsing operations, the cosine similarity between the screening results and the search conditions is calculated and sorted, and the results are displayed in combination with the display information; If the proportion is less than the preset value, a comprehensive display analysis signal is generated, the browsing operation with the largest proportion of quantity is found, the screening results are sorted by cosine similarity with the search conditions, the results are displayed, and the display results are transmitted to the display output module at the same time; The display output module is used to display the display results to the corresponding search users.

[0033] Some of the data in the above formulas are calculated by taking their numerical values, and are not substituted into parameter units for calculation. At the same time, the contents not described in detail in this specification belong to the existing technologies known to those skilled in the art.

[0034] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A method for displaying intellectual property information, characterized in that: The method specifically comprises the following steps: Obtain results and corresponding conditions based on search conditions, split search conditions and mark keywords using a large language model, filter and classify search results, and obtain classified filtering results; Select any set of classification screening results to determine duplicate data, retain the latest one as the standard result, hide the rest, and process all groups to obtain the final standard result set; Extract keywords from the screening results without duplicate data, calculate the key percentage value, sort by this value to generate sorting information, and do this for all classification screening results; Obtain sorting information and user history data, divide into segments by time T, count the percentage of user browsing operations in each segment, and compare with the preset value; If the proportion is greater than the preset value, a quantitative display analysis signal is generated and the results are displayed in combination with relevant information; if it is less than the preset value, a comprehensive display analysis signal is generated and the results are displayed as required.

2. A method for displaying intellectual property information according to claim 1, characterized in that: The specific method of obtaining the classification screening results is: Obtain the corresponding search results according to the search conditions, and at the same time obtain the search conditions corresponding to the search results, use the large language model to split the search conditions to obtain split information, and mark the keywords, and at the same time filter the search results according to the split information to generate filtered results, and at the same time classify the filtered results based on the proportion of the number of keywords in the search conditions to generate classified filtering results.

3. A method for displaying intellectual property information according to claim 1, characterized in that: The specific method of obtaining the final standard result set is: Select any group of classification and screening results as the analysis object, and obtain the search results therein, determine the duplicate data based on the similarity of the search result content, retain the latest one in the duplicate data as the standard result, and mark the rest as hidden results. All classification and screening results are processed in this way to obtain the final standard result set.

4. A method for displaying intellectual property information according to claim 1, characterized in that: The specific method of generating the sorting information is as follows: Get the filtering results that do not contain duplicate data, extract their keywords, calculate the ratio of the number of keywords in each filtering result to the total number of keywords, sum these ratios to get the key ratio value, sort the filtering results from large to small according to the key ratio value, generate sorting information, repeat this operation for all classified filtering results, and generate their respective sorting information.

5. A method for displaying intellectual property information according to claim 1, characterized in that: The specific method of comparing the statistical proportion of user browsing operations in each segment with the preset value is as follows: Obtain sorting information and retrieve user historical data, segment the historical data with time T as the period, obtain b historical period segments a, count the user browsing operations in each period, find out the operations with the same number of views, count the corresponding number of search results, calculate their proportion, and compare it with the preset number proportion set by the operator.

6. A method for displaying intellectual property information according to claim 1, characterized in that: The specific method of generating the quantitative display analysis signal and displaying the result in combination with the relevant information is as follows: If the proportion is greater than the preset value, a quantity display analysis signal is generated, display information is generated according to the same number of browsing operations, the cosine similarity between the screening results and the search conditions is calculated and sorted, and the results are displayed in combination with the display information.

7. A method for displaying intellectual property information according to claim 1, characterized in that: The specific method of generating a comprehensive display analysis signal and displaying the results as required is: If the proportion is less than the preset value, a comprehensive display analysis signal is generated, the browsing operation with the largest proportion of quantity is found, and the results are sorted by cosine similarity between the screening results and the search conditions, and displayed.

8. An intellectual property information display system, used to execute an intellectual property information display method according to any one of claims 1 to 7, characterized in that: include: The information acquisition module is used to obtain the corresponding search results according to the search conditions, and at the same time obtain the search conditions corresponding to the search results, and transmit them to the search analysis module; The retrieval analysis module is used to use the large language model to split the retrieval conditions to obtain split information and mark keywords, and at the same time, filter the retrieval results according to the split information to generate filtering results, and classify the filtering results according to the proportion of the number of keywords in the retrieval conditions to generate classified filtering results, and select a group of classified filtering results as the analysis object, and obtain the retrieval results therein, determine the duplicate data according to the similarity of the retrieval results content, retain the latest time in the duplicate data as the standard result, and mark the rest as hidden results, and process all the classified filtering results in this way to obtain the final standard result set, and transmit the final standard result set to the secondary analysis module; The secondary analysis module is used to obtain the screening results that do not contain duplicate data, extract its keywords, calculate the ratio of the number of keywords in each screening result to the total number of keywords, sum these ratios to obtain the key ratio value, sort the screening results from large to small according to the key ratio value, generate sorting information, repeat this operation for all classified screening results, and generate their respective sorting information; Obtain sorting information and retrieve user historical data, segment the historical data with time T as the period, obtain b historical period segments a, count the user browsing operations in each period, find out the operations with the same number of views, count the number of corresponding search results, calculate their proportion, and compare with the preset number proportion set by the operator; If the proportion is greater than the preset value, a quantity display analysis signal is generated, display information is generated according to the number of the same browsing operations, the cosine similarity between the screening results and the search conditions is calculated and sorted, and the results are displayed in combination with the display information; If the proportion is less than the preset value, a comprehensive display analysis signal is generated, the browsing operation with the largest proportion of quantity is found, the screening results are sorted by cosine similarity with the search conditions, the results are displayed, and the display results are transmitted to the display output module at the same time; The display output module is used to display the display results to the corresponding search users.

Images