A patent push method

Abstract

The present invention provides a patent push method and system, which relates to the field of artificial intelligence technology. The method comprises the following steps: obtaining first patent information from an enterprise user; obtaining first patent data through the first patent information; extracting patent text keywords from the first patent data; constructing patent rating rules based on the extracted patent text keywords, which are used to evaluate the importance and relevance of the patent text keywords; performing a hierarchical search on patents using the rating rules constructed using the patent text keywords to obtain a list of recommended patent data; and sorting each patent in the recommended patent data and pushing it to the enterprise user, so that the enterprise user can solve the technical problem that the direction of patent recommendation depends on expert judgment and has low efficiency, thereby achieving the technical effect of improving the automation of patent recommendation and thus improving processing efficiency.

Description

Background Art

[0002] Patent strategy has become one of the important factors to ensure the stable development of enterprises. As intangible assets, the transaction frequency between patents is increasing. How to match appropriate patents for enterprises, namely patent recommendation, has become an important research direction.

[0003] In the process of implementing the embodiments of the present application, the applicant discovered that the patent push method of the prior art is often passive push, that is, the enterprise user sends a certain demand, and then the search and push are performed based on the demand, but this push method has the problems of small business volume and low efficiency.

[0004] Although there are some active push methods in the prior art, this method often conducts a general analysis through the company's application scope, production products, and patent status. However, this method has poor pertinence because the analyzed data is relatively broad. As a result, most of the patents pushed are patents that corporate users do not need, and the push accuracy is poor. Summary of the invention

[0005] This application provides a patent push method to solve the technical problem that the existing technology is too focused on the accurate capture of patents, while the direction of patent recommendation still relies on expert judgment, resulting in low efficiency.

[0006] In view of the above problems,

[0007] This application provides a patent push method and system.

[0008] A patent push method, the method comprising:

[0009] Obtain first patent information through corporate users;

[0010] Obtaining first patent data through first patent information;

[0011] extracting patent text keywords from the first patent data;

[0012] Based on the extracted patent text keywords, patent rating rules are constructed, which are used to evaluate the importance and relevance of patent text keywords;

[0013] Using the rating rules constructed by patent text keywords, perform a hierarchical search on patents to obtain a list of recommended patent data;

[0014] The patents in the recommended patent data list are sorted and pushed to the corporate user.

[0015] Furthermore, in order to use the search results more accurately, the method of constructing patent rating rules based on the extracted patent text keywords is as follows:

[0016] The patent text keywords extracted from the patent title and independent claims are used as the first-level keywords;

[0017] The main classification number of the patent extracted from the patent classification number is used as the second-level keyword, and the patent text keywords extracted from the patent dependent claims are used as the third-level keywords.

[0018] By grading keywords, it is possible to create a favorable foundation for the generation of subsequent search formulas.

[0019] Furthermore, the method for extracting patent text keywords from patent information is as follows:

[0020] Deriving paid patent information of the enterprise user from the first patent information;

[0021] Obtain the first patent data through the paid patent information of enterprise users;

[0022] Patent text keywords are extracted from the patent information in the first patent data as first patent text keywords.

[0023] The inventors found in the process of implementing the embodiments of the present application that the existing patent recommendation method is often to have enterprise users input specific patents for retrieval, and then extract keywords based on the specific patents for analysis to obtain a recommended patent data list, but this method is prone to the following problems:

[0024] Most of the existing patent owners have no motivation to actively trade or look for related patents. They often look for relevant institutions to search, analyze and process when their own patents are in danger. This is too passive. The existing patent push is often based on the customer's products, technical fields and other directions to search for patents and actively push them. However, the patents pushed are mostly patents that corporate users do not need, and their accuracy is poor.

[0025] Therefore, in order to solve this technical problem, the inventor discovered that in the existing patent maintenance process, the payment of patent annual fees is a very important part, and existing corporate users often adopt the approach of not paying annual fees for unimportant patents. Therefore, if the patent payment information of corporate users is obtained and analyzed based on this information, the patent information required by corporate users can be obtained more accurately with a high probability. The patent payment information of corporate users includes the annual fee paid information and the annual fee to be paid information. The annual fee to be paid information refers to the valid patents that have been newly applied for but have not yet paid the annual fee.

[0026] Therefore, the first patent data can be obtained by utilizing the existing patent payment platform or a self-established platform that includes a patent payment function, through the user information of the patent payment platform and the information on the user's annual fee payment.

[0027] Furthermore, the first patent data is the patent for which the corporate user has recently paid annual fees.

[0028] Furthermore, the method for obtaining the patent information paid by the enterprise user is as follows:

[0029] Obtain the paid patent information of enterprise users through the patent annual fee payment platform;

[0030] The paid patent information is obtained as the first patent data.

[0031] The applicant found that the time when the corporate users paid for patent information is often different, and the corporate users often have a deeper impression of the patents analyzed and pushed based on the patents for which the corporate users have recently paid annual fees. Therefore, the accuracy and timeliness of the patents analyzed and pushed based on this patent are greatly improved. This method is most effective for corporate users with a small number of patents.

[0032] Furthermore, another way to obtain the first patent data is as follows:

[0033] Obtain the paid patent information of the enterprise user through the enterprise user;

[0034] Analyze the paid patent information of the enterprise user;

[0035] The core patents of the enterprise user are selected from the paid patent information of the enterprise user as the first patent data.

[0036] As a preferred method, the method of screening out the core patents of the enterprise user from the paid patent information of the enterprise user is as follows:

[0037] Acquire first sales information of an enterprise user, wherein the first sales information is sales of a first product type within a first preset time zone;

[0038] Extracting the first product types whose first sales information meets a first sales threshold and sorting them to obtain a first sorting result;

[0039] Traversing the first sorting to obtain the main product information of the enterprise user;

[0040] Screen the paid patent information of the enterprise user through the main product information to obtain the first patent collection;

[0041] Extract each key information of the first patent set and assign weights to it. Obtain the importance scores of the patents according to the weights of each key information and sort them. The patent with the highest score is used as the first patent data.

[0042] This method can be used to obtain the core patents of an enterprise through a computer, and then search around these core patents to improve the quality of push notifications while reducing the frequency of push notifications.

[0043] This method is suitable for corporate users with a large number of patents;

[0044] Furthermore, the method for extracting each key information of the first patent set and assigning weights is as follows:

[0045] The following key information of each patent information in the first patent set is extracted and weighted:

[0046] Obtaining total word count information of each patent information in the first patent set;

[0047] Obtain word count information of each independent claim in each patent information in the first patent set;

[0048] The number of dependent claims in each patent information in the first patent combination;

[0049] Weights are assigned to the above key information, wherein the weight assignment criteria for the total word count information and the number information of dependent claims is that the larger the number, the greater the weight, and the weight assignment criteria for the word count information of independent claims is that the smaller the number, the greater the weight.

[0050] Further, all patent data of the enterprise user except the first patent data is obtained through the first patent data as the second patent data;

[0051] Screening patents in the patent database according to the first patent text keywords to obtain third patent data;

[0052] Obtaining text keywords from the patent text in the third patent data as second patent text keywords;

[0053] The first patent text keywords and the second patent text keywords are merged and graded, and the intersection between the first patent text keywords and the second patent text keywords is classified as the first-level keywords, the other keywords of the first patent text except the intersection are classified as the second-level keywords, and the other keywords of the second patent text keywords and the intersection are classified as the third-level keywords.

[0054] Among them, if there are repeated keywords, they will be merged.

[0055] Furthermore, in order to further improve the accuracy of the push, the inventor adopted the following method, which is applicable to the case where the authorized patent application text and the authorized text are different:

[0056] Obtaining authorization information of the first patent data through the first patent data;

[0057] If the patent obtained through the first patent data has been authorized;

[0058] Obtain the application text and authorization text of the patent;

[0059] Extract the independent claims of the patent application text and the authorization text, and extract the patent text keywords of the independent claims of the patent application text and the authorization text;

[0060] Since the first patent data obtained in this application must be an authorized patent, the step of verifying whether the patent is authorized can be omitted. The method in this application is specifically as follows:

[0061] Obtain the application text and authorization text of the patent through the first patent data;

[0062] Extract the independent claims of the patent application text and the authorization text, and extract the patent text keywords of the independent claims of the patent application text and the authorization text;

[0063] Patent text keywords in the authorization text other than the intersection of the patent application text and the patent text keywords of the independent claims in the authorization text are used as special keywords.

[0064] In the process of implementing the embodiments of the present application, the inventor discovered that: in order to satisfy the novelty of the independent claim during the patent examination process, it is often necessary to modify the original independent claim, and the modified part of the original claim is often the main invention point of the patent. Therefore, the inventor compares the patent application text with the patent text keywords extracted from the independent claim in the authorization text, and uses the patent text keywords in the authorization text other than the intersection of the patent application text and the patent text keywords of the independent claim in the authorization text as special keywords, so as to further increase the weight of the special keywords in the process of patent search analysis and push sorting, so as to make the pushed patents more accurate.

[0065] Furthermore, the patents in the patent data list are sorted and pushed to the enterprise user, including:

[0066] Construct patent push rating rules to sort the patents in the patent data list.

[0067] Even if the patents ranked at the top of the patent recommendation list received by the enterprise user are all patents that are highly relevant to the enterprise user's products;

[0068] The method for constructing patent push rating rules to sort the patents in the patent data list is as follows:

[0069] Traverse the patent texts of each patent in the patent data list and score each major impact dimension in the patent text;

[0070] Obtain patent text keywords and assign weights to each keyword according to patent rating rules;

[0071] Calculate the matching scores of keywords in the patent text to determine the relevance score of the patent;

[0072] Obtaining technical field information from patent information, where the technical field information is the patent classification number, assigning a weight to each patent classification number based on the technical field classification of the patent, and determining the technical field score of the patent by calculating the matching score of the technical field in the patent text;

[0073] Obtain applicant information from patent information, and evaluate the patent value of applicants for each patent based on the number, type, and proportion of valid patents of the applicant to determine the influence score of the patent owner;

[0074] According to the patent application date in the patent information and the validity of the patent, weights are allocated, and then the patent validity score of each patent is determined;

[0075] The scores of each dimension are weighted and combined to obtain the total score of the patent. The patents in the patent data list are sorted according to the total scores of each patent in the patent data list and pushed to the corporate user.

[0076] As a preferred method: the method of obtaining patent text keywords and assigning weights to each keyword according to patent rating rules is as follows;

[0077] The patent text keyword weight distribution ratio is distributed in the following order from large to small;

[0078] The order is as follows: special keywords, first-level keywords, second-level keywords, and third-level keywords.

[0079] A patent push method, characterized in that the system comprises:

[0080] Patent information acquisition module, used to obtain patent information queried by enterprise users;

[0081] Keyword extraction module, used to extract patent text keywords from patent information;

[0082] A patent rating rule building module is used to build patent rating rules based on the extracted patent text keywords. These rules are used to evaluate the importance and relevance of the patent text keywords.

[0083] The patent search and analysis module uses patent text keywords and constructed rating rules to perform hierarchical searches on patents and obtain a patent data list;

[0084] The patent push module sorts the patents in the patent data list and pushes them to the corporate user.

[0085] One or more technical solutions provided by this application have at least the following technical effects or advantages:

[0086] 1. This application starts with the enterprise's patent annual fee payment information to obtain the enterprise's core patents, and uses its core patents as the basis for analysis as the first patent data, so that it can increase the accuracy of patent push and avoid the problem of too frequent patent push when the enterprise user owns a large number of patents.

[0087] 2. This application obtains the first patent data by analyzing the annual patent fee payment information of the enterprise, and further optimizes the goal of patent keyword extraction through further analysis of the first patent data. It can also be used by corporate users with fewer patents.

[0088] 3. This application has made further optimization by extracting keywords from patent information. By comparing the patent application text and the authorization text, the core special keywords are obtained for patent retrieval and sorting of search results, which further increases the matching degree between the pushed patents and the target patents.

[0089] 4. This application establishes patent push rating rules to sort the patents in the patent data list to further increase the accuracy of patent push.

[0090] 5. This application conducts patent analysis on corporate users to avoid the impact of overly complex patent data from other non-corporate users such as universities, individuals, and research institutions on the search results of this application.

[0091] 6. This application further optimizes the extraction and grading of patent keywords, so that it can obtain other related patents under the name of the enterprise user through the first patent information, and extract, compare and classify the keywords of these patents to make its search results more comprehensive.

[0092] 7. This application pushes patents through the patent annual fee payment information of enterprise users, so that it can help enterprises learn about the patent status of the enterprise's current main technologies and products, so that it can guide the innovation direction of the enterprise and avoid the enterprise from conducting research in the direction of patents applied for by others. It can also know the advanced level of its own product technology, so that the enterprise can assist in deciding the direction of research and development of the enterprise based on the patents recommended on the patent data list, further stimulate the hidden patent transaction demand of the enterprise, and further increase the transaction liquidity of the patent trading market.

[0093] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0094] Figure 1 A schematic diagram of a patent push method flow is provided for an embodiment of the present application;

[0095] Figure 2 A flowchart of another method for obtaining first patent data in a patent push method is provided for an embodiment of the present application;

[0096] Figure 3 A flowchart of a method for obtaining special keywords in a patent push method is provided for an embodiment of the present application;

[0097] Figure 4 A flowchart of a method for sorting and pushing patents in a patent data list in a patent push method is provided for an embodiment of the present application;

[0098] Figure 5 A schematic diagram of the structure of a patent push system is provided for the embodiment of the present application;

[0099] Figure 6 It is a schematic diagram of the structure of an exemplary electronic device according to an embodiment of the present application.

[0100] Explanation of the reference numerals: patent information acquisition module 10, keyword extraction module 20, patent rating rule construction module 30, patent search and analysis module 40, patent push module 50, electronic device 3, memory 301, processor 302, communication interface 303, bus architecture 304. DETAILED DESCRIPTION

[0101] The embodiment of the present application provides a patent push method and system to solve the technical problem that the prior art places too much emphasis on accurate capture of patents, while the direction of patent recommendation still relies on expert judgment, resulting in low efficiency, and achieves the technical effect of improving the degree of automation of patent recommendation and thus improving processing efficiency.

[0102] Application Overview:

[0103] Patent recommendation is an important part of patent transactions. The existing method relies on experts to study the existing patent information of enterprise users, determine the direction of patent recommendation, and then use natural language processing and machine learning technology to make patent recommendations based on the direction of patent recommendation. However, this method is relatively unstable and inefficient because the process of determining the direction of patent recommendation is more dependent on expert experience. In addition, the existing active push methods and systems have poor accuracy in pushing patents.

[0104] In response to the above technical problems, the overall idea of ​​the technical solution provided by this application is as follows:

[0105] The embodiment of the present application provides a patent push method and system. By determining the name of an enterprise, the patent data of the enterprise that pays an annual fee is collected; the core patents of the enterprise can be obtained more accurately through the analysis of the present application, and then the surrounding patents related to the core patents in the patents of the enterprise user are obtained through the analysis of its core patents, and then the core patents and surrounding patents of the enterprise user are analyzed and the key information of these patents is extracted for classification, grading, and obtaining a search formula, and the existing patents are retrieved through the search formula to obtain the patents to be pushed, and finally the patents in the patent data list are further sorted and sorted by constructing patent push rating rules, and the generated patent data list is pushed to the enterprise user.

[0106] After introducing the basic principles of the present application, various non-limiting implementation methods of the present application will be specifically described below in conjunction with the drawings in the specification.

[0107] Embodiment 1:

[0108] like Figure 1 As shown, the embodiment of the present application provides a patent push method, wherein the method includes:

[0109] S100 obtains the first patent information through corporate users;

[0110] Specifically, enterprise users include but are not limited to: enterprises in the form of legal entities, non-legal entities or public institutions; the enterprise name of the user is obtained through the enterprise user, and the above enterprise name refers to the name of the applicant used by the enterprise user when applying for a patent; patent information is obtained through the enterprise name, and the patent information refers to the public patent cluster, pre-application patent cluster and patent cluster under application for the applicant being the first enterprise name, including but not limited to: utility model patents, invention patents and appearance patents. By collecting the first patent data, it is convenient to determine the current patent layout of the first enterprise user in the later step. Comprehensive first patent data is the guarantee to ensure the accuracy of the later data analysis.

[0111] S200 obtains first patent data through first patent information;

[0112] Specifically, the first patent data is patent data of one of the patents in the first patent information.

[0113] S300 extracts patent text keywords from the first patent data. Preferably, the extraction scope of the patent text keywords is the patent title, claims and patent classification number, that is, the patent text keywords are extracted based on the patent title, claims and patent classification number in the first patent data.

[0114] Specifically: the text information of the patent can be obtained through the first patent data, and then semantic recognition is performed on each patent in the first patent data. The semantic recognition method is preferably processed by a semantic recognition module constructed based on a recurrent neural network. Since the nodes of the recurrent neural network can take into account the input information of the input end and the output information of the previous node, it is conducive to the semantic recognition of the context and can achieve more accurate semantic recognition. Preferably, multiple groups of text information-partial text keyword identification information are input into the semantic recognition module for semi-supervised training. When the module reaches convergence, the first patent data is input into the semantic recognition module in units of articles, and text keywords are output. Generally speaking: patent text keywords include but are not limited to: technical field, technical problem, application subject, involved product and other information.

[0115] The keyword extraction method can also use the following methods, or use them in combination to extract keywords from patent texts:

[0116] TF-IDF (Term Frequency-Inverse Document Frequency): Calculates the importance of each term in the patent text. The higher the TF-IDF value, the more important the term.

[0117] Topic model: such as LDA (Latent Dirichlet Allocation), identifies topics in patent texts and extracts keywords for each topic.

[0118] Patent-specific dictionaries: Use specialized dictionaries in the patent field to extract specific terms and keywords.

[0119] S400 builds patent rating rules based on the extracted patent text keywords. These rules are used to evaluate the importance and relevance of patent text keywords.

[0120] Specifically: it can judge the importance of the patent text keywords according to the types of extracted patent text keywords, as well as their positions and occurrence frequencies in the patent text, and construct patent rating rules in turn to facilitate the generation of subsequent search formulas.

[0121] S500 uses the rating rules constructed by patent text keywords to perform hierarchical searches on patents and obtain a list of recommended patent data;

[0122] Specifically, it can use the extracted keywords to search in the patent database, match relevant patents, and perform advanced searches based on the importance of the keywords given by the patent rating rules combined with Boolean logic (AND, OR, NOT) to further optimize its search results.

[0123] S600 sorts the patents in the recommended patent data list and pushes them to the enterprise user.

[0124] Specifically, it screens the search results based on their relevance and importance, and selects the most relevant patents for push. Its push channels: can push search results to customers through emails, SMS, App notifications and other channels. Push content: including patent number, title, abstract, patent text, relevance score, etc., to help customers quickly understand the search results.

[0125] Furthermore, in order to use the search results more accurately, the method of constructing patent rating rules according to the extracted patent text keywords in S400 includes:

[0126] S410 uses patent text keywords extracted from the patent name and independent claims in the first patent data as first-level keywords;

[0127] S420 uses the main classification number of the patent extracted from the patent classification number as the second-level keyword;

[0128] S430 uses patent text keywords extracted from patent dependent claims as third-level keywords.

[0129] Existing calculations are often performed by extracting keywords from the entire patent text, but the keywords obtained by this method are of varying numbers and there are too many invalid keywords, resulting in too few patent documents in the search results, and the problem of not being able to retrieve valuable patents. The keyword grading and extraction method of this application can generate a search equation according to the importance of the keywords.

[0130] Furthermore, the method for extracting patent text keywords from patent information in S300 includes:

[0131] S310 obtains the payment patent information of the enterprise user from the first patent information;

[0132] S320 obtains the first patent data through the paid patent information of the enterprise user;

[0133] S330 extracts patent text keywords from the patent information in the first patent data as first patent text keywords.

[0134] Wherein: as a preference, patent text keywords can be extracted from the patent subject information, claim information, and patent classification number information in the patent information in the first patent data as the first patent text keywords.

[0135] Specifically, the existing patent recommendation method is often to have enterprise users input specific patents for retrieval, and then extract keywords based on the specific patents for analysis to obtain the recommended patent data list. However, this method is prone to the following problems:

[0136] Most of the existing patent owners have no motivation to actively trade or look for related patents. They often look for relevant institutions to search, analyze and process when their own patents are in danger. This is too passive. The existing patent push is often based on the customer's products, technical fields and other directions to search for patents and actively push them. However, the patents pushed are mostly patents that corporate users do not need, and their accuracy is poor.

[0137] Therefore, in order to solve this technical problem, the inventor found that in the existing patent maintenance process, the payment of patent annual fees is a very important part, and existing corporate users often do not pay annual fees for unimportant patents. Therefore, if the patent information paid by corporate users is obtained and analyzed based on this information, the patent information required by corporate users can be obtained with a high probability and more accurately. The patent information paid by corporate users includes the information of paid annual fees and the information of pending annual fees. The pending annual fee information refers to the valid patents that have been newly applied for but have not yet paid the annual fees.

[0138] Among them: the acquisition of the paid patent information of the enterprise users can be obtained through the patent annual fee payment platform.

[0139] Further, there are two schemes for obtaining the first patent data in step S200;

[0140] Solution 1: The time of patent payment information of the enterprise user is often different, and the patent analyzed and pushed based on the patent for which the enterprise user has recently paid the annual fee is often more impressive to the enterprise user. Therefore, the accuracy and timeliness of the patent analyzed and pushed based on this patent are greatly improved. Therefore, the first patent data can be the patent for which the enterprise user has recently paid the annual fee, and the method of obtaining the patent for which the enterprise user has recently paid the annual fee can be obtained from a privately established patent payment platform through the annual fee payment information of the enterprise user on the platform.

[0141] Solution 2: If Figure 2 As shown, as another way to obtain the first patent data, the method includes:

[0142] S210 obtains the paid patent information of the enterprise user through the enterprise user;

[0143] S220 analyzes the payment patent information of the enterprise user;

[0144] S230 selects the core patent of the enterprise user from the paid patent information of the enterprise user as the first patent data;

[0145] Among them, the method of S230 screening out the core patents of the enterprise user from the paid patent information of the enterprise user includes:

[0146] S231: obtaining first sales information of an enterprise user, wherein the first sales information is sales of a first product type within a first preset time zone;

[0147] S232 extracts the first product types whose first sales information meets a first sales threshold and sorts them to obtain a first sorting result;

[0148] S233 traverses the first sorting to obtain the main product information of the enterprise user;

[0149] S234 screens the paid patent information of the enterprise user through the main product information to obtain a first patent set;

[0150] S235 extracts each key information of the first patent set and assigns weights to them, obtains the importance scores of the patents by the weights of each key information, and sorts them, and the patent with the highest score is used as the first patent data.

[0151] Furthermore, in step S235, the method of extracting each key information of the first patent set and assigning weights includes:

[0152] S2351 extracts the following key information of each patent information in the first patent set and assigns weights:

[0153] S2352 obtains the total number of words in each patent information in the first patent set;

[0154] S2353 obtains the word count information of each independent claim in each patent information in the first patent combination;

[0155] Information on the number of dependent claims in each patent information in the first patent combination obtained by S2354

[0156] S2355 assigns weights to the above key information, wherein the weight assignment standard for the total word count information and the number information of the dependent claims is that the larger the number, the greater the weight, and the weight assignment standard for the word count information of the independent claim is that the smaller the number, the greater the weight.

[0157] Specifically, its weight distribution can be adjusted and added according to the above rules. Its total number of words often represents the degree of attention paid to the patent technology and the degree of detail of its disclosure. Generally speaking, the more attention the applicant pays to the patent, the more words its patent has, and the more difficult it is to invalidate. From the side, it can be concluded that the value and writing quality of the patent are often positively correlated with the number of words in the patent. The fewer words an independent claim has in the authorization text, the fewer technical features the claim has, that is, the scope of protection of the patent is more considerable, and the more dependent rights there are, the stronger the patent's ability to resist risks. Therefore, by comprehensively evaluating these three aspects, the user's core patent can often be obtained, and analysis based on this can greatly improve the effect of patent push.

[0158] Furthermore, step S300 also includes:

[0159] S340 obtains all patent data of the enterprise user except the first patent data as second patent data through the first patent data;

[0160] S350 screens patents in the patent database according to the first patent text keyword to obtain third patent data;

[0161] S360 obtains a text keyword from the patent text in the third patent data as a second patent text keyword;

[0162] S370 merges and grades the first patent text keywords and the second patent text keywords, classifies the intersection between the first patent text keywords and the second patent text keywords into first-level keywords, classifies other keywords of the first patent text keywords except the intersection into second-level keywords, and classifies other keywords of the second patent text keywords and the intersection into third-level keywords.

[0163] Specifically: Performing a hierarchical search on patents to obtain recommended patent data, the method example is as follows;

[0164] Identify Keywords: Identify all keywords and categorize them into first-level, second-level, and third-level keywords.

[0165] Use natural language processing (NLP) tools or synonyms to automatically obtain synonyms and near-anonyms.

[0166] Expand the keyword list:

[0167] Expand the keyword list at each level:

[0168] Add the found synonyms and near-synonyms to the corresponding keyword list.

[0169] For example:

[0170] First-level keywords: glass (synonyms: glass material) lens (synonyms: lens, mirror) intelligence (synonyms: automation, wisdom) control (synonyms: control system, adjustment) compensation (synonyms: correction, correction);

[0171] Secondary Keywords: Processing (Synonyms: manufacturing, processing) Parameter (Synonyms: parameter value, parameter set) Surface (Synonyms: curve surface, curved surface) G05B19 / 404 (no synonyms, remain unchanged);

[0172] Tertiary Keywords: model (synonyms: modeling, template) adjustment (synonyms: calibration, regulation) node (synonyms: node state, point) BP neural network (synonyms: back propagation neural network);

[0173] Constructing the expanded retrieval equation;

[0174] Construct the extended retrieval equation for each level:

[0175] Connect the original keyword and its synonyms and near-synonyms with the Boolean logic OR operator.

[0176] For example:

[0177] The extended search formula for the first-level keywords is:

[0178] (glass OR glass material) AND (lens OR lens OR mirror) AND (intelligence OR automation OR wisdom) AND (control OR control system OR adjustment) AND (compensation OR correction OR correction)

[0179] The extended search formula for the second-level keywords is:

[0180] (processing OR manufacturing OR processing) AND (parameter OR parameter value OR parameter set) AND (surface OR curved surface OR curved surface) AND G05B19 / 404;

[0181] The extended search formula for the second-level keywords is:

[0182] (model OR modeling OR template) AND (adjust OR calibrate OR adjust) AND (node ​​OR node state OR point) AND (BP neural network OR back propagation neural network)

[0183] Combine the retrieval equations for all levels:

[0184] Combine the expansion keywords of each level and connect different levels with the AND operator.

[0185] Comprehensive example:

[0186] ((glass OR glass material) AND (lens OR lens OR mirror) AND (intelligent OR automation OR wisdom) AND (control OR control system OR adjustment) AND (compensation OR correction OR calibration))AND((processing OR manufacturing OR processing) AND (parameter OR parameter value OR parameter set) AND (surface OR curved surface OR curved surface)AND G05B19 / 404)AND((model OR modeling OR template) AND (adjustment OR calibration OR adjustment) AND(node ​​OR node status OR point) AND (BP neural network OR back propagation neural network))

[0187] Apply the expanded search equation:

[0188] The constructed extended search equation is applied to the patent database or information retrieval system to obtain relevant patent or literature results. This method can increase the coverage of the search results and improve the comprehensiveness and accuracy of the search results.

[0189] Through these steps, a retrieval formula that meets the needs can be effectively constructed based on the importance of keywords.

[0190] Furthermore, in order to further improve the accuracy of the push, step S300 also includes:

[0191] S321 obtains the application text and authorization text of the patent through the first patent data;

[0192] S322 extracts the independent claims of the patent application text and the authorization text, and extracts patent text keywords of the independent claims of the patent application text and the authorization text;

[0193] S323 regards the patent text keywords in the authorization text other than the intersection of the patent application text and the patent text keywords of the independent claims in the authorization text as special keywords.

[0194] Specifically: In order to meet the novelty of the independent claim during the patent examination process, the original independent claim often needs to be modified, and the modified part of the original claim is often the main invention of the patent. Therefore, the inventor compares the patent application text with the patent text keywords extracted from the independent claim in the authorization text, and uses the patent text keywords in the authorization text other than the intersection of the patent application text and the patent text keywords of the independent claim in the authorization text as special keywords, so as to further increase the weight of its special keywords in the process of patent search analysis and push ranking, so as to make the pushed patents more accurate.

[0195] Furthermore, S600 sorts the patents in the patent data list and pushes them to the enterprise user, including:

[0196] S610 constructs patent push rating rules to allocate weights to the main impact dimensions;

[0197] S620 traverses the patent texts of each patent in the patent data list, and scores each major impact dimension in the patent text according to the patent push rating rules;

[0198] S630 performs weighted integration of the scores of each dimension to obtain the total score of the patent, sorts the patents in the patent data list according to the total score of the patent, and pushes them to the corporate user.

[0199] Among them: S611 traverses the patent texts of each patent in the patent data list and scores each major impact dimension in the patent text including;

[0200] S6111 Keyword relevance dimension: Obtain patent text keywords and assign weights to keywords in each category according to patent push rating rules;

[0201] S6112 calculates the matching scores of keywords in the patent text to determine the relevance score of the patent;

[0202] S6113 Technical field dimension: obtain the technical field information in the patent information, where the technical field information is the patent classification number. According to the technical field classification of the patent, a weight is assigned to each patent classification number. The technical field score of the patent is determined by calculating the matching score of the technical field in the patent text;

[0203] S6114 Applicant influence dimension: Obtain applicant information from patent information, and evaluate the patent value of applicants of various patents based on the number, type and proportion of valid patents of the applicant to determine the influence score of the patent owner;

[0204] S6115 Patent validity dimension: Weights are assigned based on the patent application date in the patent information and the validity of the patent, and then the patent validity score of each patent is determined.

[0205] Specifically: the method of obtaining patent text keywords and assigning weights to each keyword according to patent rating rules is as follows;

[0206] The patent text keyword weight distribution ratio is distributed in the following order from large to small;

[0207] The order is as follows: special keywords, first-level keywords, second-level keywords, and third-level keywords;

[0208] Exemplary:

[0209] Weighting criteria:

[0210] Special keywords: weight 12

[0211] First-level keywords: weight 7

[0212] Second level keywords: weight 4

[0213] Third level keyword: weight 2

[0214] Rating example:

[0215] Relevance score:

[0216] Patent document A in the search results contains the following keywords:

[0217] Glass, intelligence, control, compensation, processing, parameters, model, adjustment, BP neural network, among which control and compensation are special keywords

[0218] Score calculation:

[0219] Control, compensation (2, special keywords): 2*12 = 22

[0220] Glass, Smart (2, first-level keywords): 2*7 = 14

[0221] Processing, parameters (2, second-level keywords): 2*4 = 8

[0222] Model, adjustment, BP neural network (3, third-level keywords): 3*2 = 6

[0223] Total score = 22+14+8+6 = 50.

[0224] Technical area score:

[0225] Weighting criteria:

[0226] Weight distribution is performed according to the level of classification number:

[0227] Department: Weight 1

[0228] Division: Weight 2

[0229] Category: 5

[0230] Subcategories: 7

[0231] Large group: 10

[0232] Group: 15

[0233] If the classification numbers are exactly the same, the total score = 1+2+5+7+10+15=40.

[0234] Applicant Impact Score:

[0235] Examples of the number and weight of valid invention patents of its applicants are as follows:

[0236] The weight of each invention patent is 3

[0237] The weight of each utility model patent is 2

[0238] That is, the applicant of this patent has 10 invention patents and 5 utility model patents, so his total score = 3*10+2*5=40.

[0239] Effectiveness score:

[0240] The weight of the remaining valid years of the patent is 2

[0241] That is, if the patent has 15 years of protection left after paying the annual fee, the total score of the patent is 30;

[0242] The patent is newly applied for and in valid status, with a score of 7.

[0243] Total score calculation:

[0244] Relevance score: Weight 1

[0245] Technical field score: weight 0.8

[0246] Applicant’s influence score: weight 0.5

[0247] Effectiveness score: Weight 0.2

[0248] Total score = (50*1)+(40*0.8)+(40*0.5)+(40*0.2)=110.

[0249] Sorting and pushing:

[0250] The patents in the patent data list are scored and sorted according to the total score, and the sorted patent data list is pushed to the enterprise user.

[0251] Through the above steps, each patent in the patent data list can be effectively scored and sorted, so as to push the most relevant and valuable patents to corporate users. This method ensures the relevance of patents, the matching degree of technical fields, the influence of applicants and the validity of patents, thereby improving the accuracy and practicality of pushing patents.

[0252] Embodiment 2:

[0253] Based on the same inventive concept as a patent push method in the aforementioned embodiment, Figure 3 As shown, the embodiment of the present application provides a patent push method, wherein the system includes:

[0254] A patent push method, characterized in that the system comprises:

[0255] Patent information acquisition module, used to obtain patent information queried by enterprise users;

[0256] Keyword extraction module, used to extract patent text keywords from patent information;

[0257] A patent rating rule building module is used to build patent rating rules based on the extracted patent text keywords. These rules are used to evaluate the importance and relevance of the patent text keywords.

[0258] The patent search and analysis module uses patent text keywords and constructed rating rules to perform hierarchical searches on patents and obtain a patent data list;

[0259] The patent push module sorts the patents in the patent data list and pushes them to the corporate user.

[0260] The patent rating rule building module also includes:

[0261] Using patent text keywords extracted from the patent name in the first patent data and the independent claims as first-level keywords;

[0262] The main classification number of the patent extracted from the patent classification number is used as the second-level keyword, and the patent text keywords extracted from the patent dependent claims are used as the third-level keywords.

[0263] The keyword extraction module also includes:

[0264] Deriving paid patent information of the enterprise user from the first patent information;

[0265] Obtain the first patent data through the paid patent information of enterprise users;

[0266] Patent text keywords are extracted from the patent information in the first patent data as first patent text keywords.

[0267] There are two schemes for obtaining the first patent data in the patent information acquisition module;

[0268] Solution 1: The time when the enterprise user pays for patent information is often different. The enterprise user often has a deeper impression of the patent analyzed and pushed based on the patent for which the enterprise user has recently paid the annual fee. Therefore, the accuracy and timeliness of the patent analyzed and pushed based on this patent are greatly improved. Therefore, the first patent data can be the patent for which the enterprise user has recently paid the annual fee.

[0269] Solution 2: Another method for obtaining the first patent data includes:

[0270] Obtain the paid patent information of the enterprise user through the enterprise user;

[0271] Analyze the paid patent information of the enterprise user;

[0272] Filtering the core patents of the enterprise user from the paid patent information of the enterprise user as the first patent data;

[0273] The method of selecting the core patents of the enterprise user from the paid patent information of the enterprise user includes:

[0274] Acquire first sales information of an enterprise user, wherein the first sales information is sales of a first product type within a first preset time zone;

[0275] Extracting the first product types whose first sales information meets a first sales threshold and sorting them to obtain a first sorting result;

[0276] Traversing the first sorting to obtain the main product information of the enterprise user;

[0277] Screen the paid patent information of the enterprise user through the main product information to obtain the first patent collection;

[0278] Extract each key information of the first patent set and assign weights to it. Obtain the importance scores of the patents according to the weights of each key information and sort them. The patent with the highest score is used as the first patent data.

[0279] The method for extracting the key information of the first patent set and assigning weights in the step includes:

[0280] The following key information of each patent information in the first patent set is extracted and weighted:

[0281] Obtaining total word count information of each patent information in the first patent set;

[0282] Obtain word count information of each independent claim in each patent information in the first patent combination;

[0283] The number of dependent claims in each patent information in the first patent combination;

[0284] Weights are assigned to the above key information, wherein the weight assignment criteria for the total word count information and the number information of dependent claims is that the larger the number, the greater the weight, and the weight assignment criteria for the word count information of independent claims is that the smaller the number, the greater the weight.

[0285] Among them; the keyword extraction module also includes:

[0286] Obtain all patent data of the enterprise user except the first patent data through the first patent data as the second patent data;

[0287] Screening patents in the patent database according to the first patent text keywords to obtain third patent data;

[0288] Obtaining text keywords from the patent text in the third patent data as second patent text keywords;

[0289] The first patent text keywords and the second patent text keywords are merged and graded, and the intersection between the first patent text keywords and the second patent text keywords is classified as the first-level keywords, the other keywords of the first patent text except the intersection are classified as the second-level keywords, and the other keywords of the second patent text keywords and the intersection are classified as the third-level keywords.

[0290] In order to further improve the accuracy of push notifications, the keyword extraction module also includes:

[0291] Obtaining authorization information of the first patent data through the first patent data;

[0292] If the patent obtained through the first patent data has been authorized;

[0293] Obtain the application text and authorization text of the patent;

[0294] Extract the independent claims of the patent application text and the authorization text, and extract the patent text keywords of the independent claims of the patent application text and the authorization text;

[0295] Patent text keywords in the authorization text other than the intersection of the patent application text and the patent text keywords of the independent claims in the authorization text are used as special keywords.

[0296] The patent push module also includes sorting the patent data and pushing it to the enterprise user, including:

[0297] Construct patent push rating rules to sort the patents in the patent data.

[0298] Among them: the method of constructing patent push rating rules to sort the patents in the patent data includes:

[0299] Traverse the patent texts of each patent in the patent data and score each major impact dimension in the patent text;

[0300] Obtain patent text keywords and assign weights to each keyword according to patent rating rules;

[0301] Calculate the matching scores of keywords in the patent text to determine the relevance score of the patent;

[0302] Obtaining technical field information from patent information, where the technical field information is the patent classification number, assigning a weight to each patent classification number based on the technical field classification of the patent, and determining the technical field score of the patent by calculating the matching score of the technical field in the patent text;

[0303] Obtain applicant information from patent information, and evaluate the patent value of applicants for each patent based on the number, type, and proportion of valid patents of the applicant to determine the influence score of the patent owner;

[0304] According to the patent application date in the patent information and the validity of the patent, weights are allocated, and then the patent validity score of each patent is determined;

[0305] The scores of each dimension are weighted and combined to obtain the total score of the patent. The patents in the patent data are sorted according to the total score and pushed to the corporate user.

[0306] Specifically: the method of obtaining patent text keywords and assigning weights to each keyword according to patent rating rules is as follows;

[0307] The patent text keyword weight distribution ratio is distributed in the following order from large to small;

[0308] The order is as follows: special keywords, first-level keywords, second-level keywords, and third-level keywords.

[0309] Example electronic devices:

[0310] Reference below Figure 3 To describe the electronic device of the embodiment of the present application.

[0311] Based on the same inventive concept as the patent recommendation method based on enterprise portrait in the aforementioned embodiment, the embodiment of the present application also provides an electronic device, including: a processor, the processor is coupled to a memory, the memory is used to store a program, when the program is executed by the processor, the processor executes the program to implement any one of the methods described in embodiment one.

[0312] The electronic device 30 includes: a processor 302, a communication interface 303, and a memory 301. Optionally, the electronic device 300 may also include a bus architecture 304. The communication interface 303, the processor 302, and the memory 301 may be interconnected via the bus architecture 304; the bus architecture 304 may be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus. The bus architecture 304 may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 4 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0313] The processor 302 may be a CPU, a microprocessor, an ASIC, or one or more integrated circuits for controlling the execution of the program of the present application.

[0314] The communication interface 303 uses any transceiver-like system for communicating with other devices or communication networks, such as Ethernet, radio access network (RAN), wireless local area networks (WLAN), wired access networks, etc.

[0315] The memory 301 may be a ROM or other type of static storage device that can store static information and instructions, a RAM or other type of dynamic storage device that can store information and instructions, or an electrically erasable programmable read-only memory (EEPROM), a compact disc (CD-ROM) or other optical disc storage, an optical disc storage (including a compressed optical disc, a laser disc, an optical disc, a digital versatile disc, a Blu-ray disc, etc.), a magnetic disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store the desired program code in the form of an instruction or data structure and can be accessed by a computer, but is not limited thereto. The memory may exist independently and be connected to the processor via a bus architecture 304. The memory may also be integrated with the processor.

[0316] The memory 301 is used to store computer execution instructions for executing the solution of the present application, and the execution is controlled by the processor 302. The processor 302 is used to execute the computer execution instructions stored in the memory 301, thereby realizing a patent recommendation method based on enterprise portrait provided in the above embodiment of the present application.

[0317] Optionally, the computer-executable instructions in the embodiments of the present application may also be referred to as application code, which is not specifically limited in the embodiments of the present application.

[0318] The embodiment of the present application provides a patent push method and system. Due to the adoption of the method of determining the enterprise name, collecting patent data under the enterprise name; then collecting the sales information of various products of the enterprise; sorting the products that meet the sales threshold; matching the patent data with the corresponding products one by one according to the sorting results, filtering out the remaining patent data, and obtaining the screened data; stratifying the screened data to obtain three levels of patents with decreasing importance; performing defect analysis on the three levels of patents in turn to obtain defect information; constructing the enterprise patent portrait according to the sorting results and the three levels of patents; determining the patent recommendation label according to the defect information, and adding it to the technical solution of the enterprise patent portrait, the defect analysis is performed on the relevant patent data of the enterprise's products with higher sales, that is, the core products, to construct the enterprise patent portrait, and then realize the automatic confirmation of the recommended patents for the defect information, so as to achieve the technical effect of improving the degree of automation of patent recommendation and thus improving the processing efficiency.

[0319] Those skilled in the art will appreciate that the various digital numbers such as the first and second involved in the present application are only for the convenience of description, and are not used to limit the scope of the embodiments of the present application, nor do they represent the order of precedence. "And / or" describes the association relationship of the associated objects, indicating that three relationships may exist. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. The character " / " generally indicates that the associated objects before and after are in an "or" relationship. "At least one" refers to one or more. At least two refers to two or more. "At least one", "any one" or similar expressions refer to any combination of these items, including any combination of single items (individuals) or plural items (individuals). For example, at least one item (individual, kind) of a, b, or c can represent: a, b, c, a-b, ac, bc, or abc, where a, b, c can be single or multiple.

[0320] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable systems. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium, for example, the computer instructions can be transmitted from a website site, computer, server or data center by wired (e.g., coaxial cable, optical fiber, digital enterprise subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) mode to another website site, computer, server or data center. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server, a data center, etc. that includes one or more available media integrated. The available medium may be a magnetic medium (eg, a floppy disk, a hard disk, a magnetic tape), an optical medium (eg, a DVD), or a semiconductor medium (eg, a solid state disk (SSD)).

[0321] The various illustrative logic units and circuits described in the embodiments of the present application can be implemented or operated by a general-purpose processor, a digital signal processor, an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic system, a discrete gate or transistor logic, a discrete hardware component, or the design of any combination of the above functions. The general-purpose processor can be a microprocessor, and optionally, the general-purpose processor can also be any traditional processor, controller, microcontroller or state machine. The processor can also be implemented by a combination of a computing system, such as a digital signal processor and a microprocessor, a plurality of microprocessors, one or more microprocessors in conjunction with a digital signal processor core, or any other similar configuration to implement.

[0322] The steps of the method or algorithm described in the embodiments of the present application can be directly embedded in hardware, a software unit executed by a processor, or a combination of the two. The software unit can be stored in a RAM memory, a flash memory, a ROM memory, an EPROM memory, an EEPROM memory, a register, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium in the art. Exemplarily, the storage medium can be connected to the processor so that the processor can read information from the storage medium and can write information to the storage medium. Optionally, the storage medium can also be integrated into the processor. The processor and the storage medium can be arranged in an ASIC, and the ASIC can be arranged in a terminal. Optionally, the processor and the storage medium can also be arranged in different components in the terminal. These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are performed on the computer or other programmable device to produce a computer-implemented process, so that the instructions executed on the computer or other programmable device provide for implementing the process in the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0323] Although the present application has been described in conjunction with specific features and embodiments thereof, it is obvious that various modifications and combinations may be made thereto without departing from the spirit and scope of the present application. Accordingly, this specification and the accompanying drawings are merely exemplary illustrations of the present application as defined therein, and are deemed to have covered any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, a person skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application is intended to include these modifications and variations.

Claims

1. A patent push method, characterized in that: The method comprises: Obtain first patent information through corporate users; Obtaining first patent data through first patent information, where the first patent data is a patent for which the enterprise user has recently paid an annual fee; extracting patent text keywords from the first patent data; Based on the extracted patent text keywords, patent rating rules are constructed. These rules are used to evaluate the importance and relevance of patent text keywords. Using the rating rules constructed by patent text keywords, perform a hierarchical search on patents to obtain a list of recommended patent data; sorting the patents in the recommended patent data list and pushing them to the enterprise user; Among them, the method of constructing patent rating rules based on the extracted patent text keywords is as follows: The patent text keywords extracted from the patent title and independent claims are used as the first-level keywords; The main classification number of the patent extracted from the patent classification number is used as the second-level keyword; The patent text keywords extracted from the patent dependent claims are used as the third-level keywords; Deriving paid patent information of the enterprise user from the first patent information; Obtain the first patent data through the paid patent information of enterprise users; Extracting patent text keywords from patent information in the first patent data as first patent text keywords; Obtain all patent data of the enterprise user except the first patent data through the first patent data as the second patent data; Screening patents in the patent database according to the first patent text keywords to obtain third patent data; Obtaining text keywords from the patent text in the third patent data as second patent text keywords; Merge and grade the first patent text keyword and the second patent text keyword, classify the intersection of the first patent text keyword and the second patent text keyword into the first-level keyword, classify the other keywords of the first patent text keyword except the intersection into the second-level keyword, and classify the other keywords of the second patent text keyword and the intersection into the third-level keyword; Obtain the application text and authorization text of the patent through the first patent data; Extract the independent claims of the patent application text and the authorization text, and extract the patent text keywords of the independent claims of the patent application text and the authorization text; Patent text keywords in the authorization text other than the intersection of the patent application text and the patent text keywords of the independent claims in the authorization text are used as special keywords.

2. A patent push method according to claim 1, characterized in that: The patent data list is sorted and pushed to the enterprise user, including: Construct patent push rating rules to sort the patents in the patent data list.

3. A patent push method according to claim 2, characterized in that: The method of constructing patent push rating rules to sort the patents in the patent data list and then push them to the enterprise user is as follows: Construct patent push rating rules to allocate weights to the main impact dimensions; Traverse the patent texts of each patent in the patent data list, and score the main impact dimensions in the patent text according to the patent push rating rules, where the main impact dimensions include: keyword relevance dimension, technical field dimension, applicant influence dimension, and patent validity dimension. The keyword relevance dimension is to obtain patent text keywords and assign weights to keywords of each category according to the patent push rating rules; The scores of each dimension are weighted and combined to obtain the total score of the patent. The patents in the patent data list are sorted according to the total score of the patent and then pushed to the enterprise user; Traverse the patent texts of each patent in the patent data list and score each major impact dimension in the patent text; The scores of each dimension are weighted and combined to obtain the total score of the patent. The patents in the patent data list are sorted according to the total scores of each patent in the patent data list and pushed to the corporate user.

4. A patent push method according to claim 3, characterized in that: The method of obtaining patent text keywords and assigning weights to each keyword according to patent rating rules is as follows: The patent text keyword weight distribution ratio is distributed in the following order from large to small; The order is as follows: special keywords, first-level keywords, second-level keywords, and third-level keywords.

5. A patent push system, the system is used to execute the method described in any one of claims 1 to 4, characterized in that: The system comprises: Patent information acquisition module, used to obtain patent information queried by enterprise users; Keyword extraction module, used to extract patent text keywords from patent information; A patent rating rule building module is used to build patent rating rules based on the extracted patent text keywords. These rules are used to evaluate the importance and relevance of the patent text keywords. The patent search and analysis module uses patent text keywords and constructed rating rules to perform hierarchical searches on patents and obtain a patent data list; The patent push module sorts the patent data list and pushes it to the corporate user.

Images