Policy analysis system and method based on intelligent semantic recognition

Abstract

The present invention relates to the field of data processing technology, specifically to a policy parsing system and method based on intelligent semantic recognition, the system comprising a policy acquisition module for acquiring policy text; a project parsing module for parsing project information according to the policy text; the project parsing module comprising a semantic recognition module, a splitting module and a parsing module; the semantic recognition module is used to perform semantic recognition on the policy text; the splitting module is used to split the policy text according to the semantic recognition result to obtain policy fragments; the parsing module is used to parse the policy fragments, the parsing module stores a project information recognition data table, the parsing module obtains corresponding recognition data according to the type of the policy fragment, the recognition data includes parsing rules, and the parsing module extracts corresponding project information according to the parsing rules. A policy parsing system and method based on intelligent semantic recognition of the present application can acquire project policies and perform accurate parsing, which is convenient for enterprises to quickly build project databases.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a policy analysis system and method based on intelligent semantic recognition. Background Art

[0002] In order to promote and support enterprises and institutions to carry out technological innovation, industrial upgrading and development, improve the core competitiveness of the industry, enhance the independent innovation ability of enterprises, improve the financing environment, and achieve social and economic development goals, relevant departments at all levels usually set up corresponding special projects. However, at present, most small and micro enterprises and technology-based enterprises often fail to realize or often ignore the release of relevant policy information for various reasons, miss the opportunity to apply for projects that meet the actual situation of the enterprise itself, and at the same time fail to achieve the effect of helping the development of enterprise innovation ability and industrial competitiveness.

[0003] Some companies or consulting agencies will establish relevant project databases to manage different project information, so that they can know which projects are suitable for the company to apply for. At present, the monitoring and analysis of relevant projects and policies in the project database are usually manually operated by company personnel, which requires relevant personnel to continuously check the policies of various departments and manually record relevant project information. However, due to the large number and types of projects, manual operation has high labor costs, is prone to omissions or errors, and has poor timeliness. Summary of the invention

[0004] The present invention aims to provide a policy analysis system and method based on intelligent semantic recognition, which can obtain project policies and perform accurate analysis, so as to facilitate enterprises to quickly build project databases.

[0005] This application provides the following technical solutions:

[0006] Policy analysis system based on intelligent semantic recognition, including:

[0007] Policy acquisition module, used to obtain policy text;

[0008] A project analysis module analyzes project information according to the policy text; the project information includes: basic project information, project requirement information and project benefit information;

[0009] The project parsing module includes a semantic recognition module, a splitting module and a parsing module;

[0010] The semantic recognition module is used to perform semantic recognition on the policy text;

[0011] The splitting module is used to split the policy text according to the semantic recognition result to obtain policy fragments of various types;

[0012] The parsing module is used to parse the policy fragment according to the type of the policy fragment. The parsing module stores a project information identification data table. The project information identification data table stores a number of identification data, each identification data corresponds to a type of project information. The parsing module obtains the corresponding identification data according to the type of the policy fragment. The identification data includes parsing rules. The parsing module extracts the corresponding project information according to the parsing rules.

[0013] In the technical solution of the present invention, the project policy texts issued by various departments are automatically obtained through the policy acquisition module, and the policy texts are identified and parsed through the project parsing module using semantic parsing technology, and the project information of each project is generated based on the analysis results. The policy parsing system based on intelligent semantic recognition of the present invention can automatically acquire and parse policies to form project information, which solves the problems of existing manual operation with large workload, high cost, easy errors, and untimely updates, and can help enterprises quickly establish a project information database.

[0014] Furthermore, the project information also includes other auxiliary information, and the project parsing module also includes other auxiliary information parsing modules, and the other auxiliary information parsing modules include a key name parsing module and a key value parsing module, and the key value parsing module extracts key values ​​according to preset parsing rules, and the key name parsing module obtains corresponding key names according to semantic recognition results and the position of key values ​​in policy texts.

[0015] By automatically extracting key names and key values, the content that is not in the project information identification data table can also be identified and extracted to form other auxiliary information to avoid information omission.

[0016] Furthermore, the project requirement information includes time requirements, subject requirements and application method requirements. The subject requirements include field requirements, regional requirements, financial requirements, intellectual property requirements, establishment time, personnel requirements and qualification requirements.

[0017] Furthermore, it also includes a preprocessing module to perform formatting preprocessing on the acquired policy text data, wherein the formatting preprocessing includes removing stop words, deduplication, and garbled code correction.

[0018] Reduce the amount of data processing and improve recognition efficiency and accuracy.

[0019] Furthermore, it also includes a recognition model storage module and a model matching module, wherein the recognition model storage module is used to generate and store the recognition model according to the data source of the policy text, the splitting result of the splitting module, and the parsing result of the parsing module;

[0020] The model matching module is used to match the identification model according to the data source;

[0021] The model matching module includes a model verification module, which is used to verify whether the recognition model is valid according to the semantic recognition result;

[0022] The splitting module and the parsing module are further used to split and parse the policy text according to the effective recognition model based on the recognition module.

[0023] Each data source or website has its own customary writing format, which can be quickly split and identified by identifying templates.

[0024] Furthermore, it also includes a project information identification data table expansion module, which is used to generate identification data and add it to the project identification data table based on the analysis results of other auxiliary information analysis modules.

[0025] Furthermore, the project information identification data table expansion module includes a parsing rule generation module, and the parsing rule generation module includes a basic rule generation module and an expansion module. The basic rule generation module is used to generate basic parsing rules based on key values ​​and key names, and the expansion module is used to expand the key values ​​and key names. The expansion module is also used to expand the basic parsing rules according to the expansion results of the key values ​​and key names to form parsing rules.

[0026] Realize automatic addition and expansion of project information data table.

[0027] Furthermore, the project information identification data table expansion module also includes a standardization processing module, which performs standardization processing on each information key name according to a standardized vocabulary library.

[0028] Furthermore, it also includes a project version matching module, a comparison and analysis module and a suggestion generation module. The project version matching module is used to match the project historical version according to the project basic information, the comparison and analysis module is used to analyze the project changes according to the project historical version, and the suggestion generation module is used to generate change precautions and auxiliary suggestions according to the project changes.

[0029] Help users quickly understand project changes.

[0030] Furthermore, the present invention also discloses a policy analysis method based on intelligent semantic recognition, which uses the above-mentioned policy analysis system based on intelligent semantic recognition. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 This is a logical block diagram of an embodiment of a policy analysis system and method based on intelligent semantic recognition in the present application. DETAILED DESCRIPTION

[0032] The technical solution of this application is further described in detail below through specific implementation methods:

[0033] Embodiment 1

[0034] like Figure 1 As shown, a policy analysis system based on intelligent semantic recognition in this embodiment includes a policy acquisition module and a project analysis module.

[0035] The policy acquisition module is used to obtain policy texts; in this embodiment, the policy acquisition module includes a third-party data acquisition module and a web page data crawling module. The third-party data acquisition module is used to connect to a third-party data platform and retrieve the policy text or policy data stored therein according to the API of the third-party data platform. The web page data crawling module internally maintains several web page lists, which include information such as web page addresses and page categories. The web page data crawling module is used to crawl corresponding web page content according to the web page list and obtain the policy text. In this embodiment, the distributed technology based on Scrapy and Redis is used to deploy the web page data crawling module to form a distributed crawler network.

[0036] The project parsing module is used to parse project information according to the policy text; project information includes: basic project information, project requirement information, project benefit information and other auxiliary information; basic project information includes project name, sponsoring department, relevant contact information, etc. The project benefit information includes benefit content, collection method and collection time, etc.; project requirement information includes time requirements, subject requirements and application method requirements. Time requirements include start and end time requirements, material submission time requirements, review result time requirements and modification time requirements, etc. Subject requirements include field requirements, regional requirements, financial requirements, intellectual property requirements, establishment time, personnel requirements and qualification requirements, etc. Application method requirements include application document content requirements, application template, whether to apply online, whether to apply in paper form, application address, etc. Other auxiliary information mainly refers to other data content useful for project application that is not included in the basic project information, project requirement information and project benefit information, such as contact number, email address, precautions and other information.

[0037] The project parsing module includes a preprocessing module, a semantic recognition module, a splitting module, a parsing module, a recognition model storage module and a model matching module.

[0038] The preprocessing module is used to format the acquired policy text data. Formatting preprocessing includes removing stop words, deduplication, and garbled code correction. The above processing can reduce the amount of data processing on the one hand, and remove interference on the other hand, thereby improving the efficiency and accuracy of semantic recognition.

[0039] The semantic recognition module is used to perform semantic recognition on the policy text. In this embodiment, a semantic recognition algorithm based on the LSTM-CRF neural network model is used to perform word segmentation, named entity recognition, part-of-speech tagging, sentence and paragraph keyword extraction and tagging of the policy text.

[0040] The splitting module is used to split the policy text according to the semantic recognition results to obtain policy fragments of various types; specifically, in this implementation, the corresponding type of each sentence is determined according to the keywords appearing in each sentence. These types represent the type of project information protected by each policy fragment, namely, basic project information, project requirement information, project benefit information or other auxiliary information.

[0041] The parsing module is used to parse the policy fragment according to the type of the policy fragment. The parsing module stores a project information identification data table. The project information identification data table stores a number of identification data. Each identification data corresponds to a type of project information. Each identification data has its applicable policy fragment type. The parsing module obtains the corresponding identification data according to the type of the policy fragment. The identification data includes parsing rules. The parsing module extracts the corresponding project information according to the parsing rules. The parsing rules include one or more of parsing keywords, parsing ranges, regular expressions, etc.

[0042] The project parsing module also includes other auxiliary information parsing modules, which include key name parsing modules and key value parsing modules. The key value parsing module extracts key values ​​according to preset parsing rules. In this implementation, the preset parsing rules include matching according to fixed types or formats or regular expressions, such as telephone format, email format, etc. The key name parsing module obtains the corresponding key name based on the semantic recognition result and the position of the key value in the policy text. For example, when an email address is obtained, it searches forward to determine the name or related expression corresponding to the email address.

[0043] The recognition model storage module generates and stores the recognition model according to the data source of the policy text, the splitting results of the splitting module and the parsing results of the parsing module; the model matching module is used to match the recognition model according to the data source; the model matching module includes a model verification module, which is used to verify whether the recognition model is valid according to the semantic recognition results; the splitting module and the parsing module are also used to split and parse the policy text according to the effective recognition model based on the recognition module.

[0044] It also includes a project information identification data table expansion module, which is used to generate identification data and add it to the project identification data table based on the analysis results of other auxiliary information analysis modules.

[0045] The project information identification data table expansion module includes a parsing rule generation module, which includes a basic rule generation module and an expansion module. The basic rule generation module is used to generate basic parsing rules based on key values ​​and key names, and the expansion module is used to expand key values ​​and key names. The expansion module is also used to expand the basic parsing rules according to the expansion results of key values ​​and key names to form parsing rules.

[0046] The project information identification data table expansion module also includes a standardization processing module, which standardizes each information key name according to a standardized vocabulary library, such as "start time", "start time", "start time", etc., which are uniformly standardized as the start time.

[0047] It also includes a project version matching module, a comparative analysis module and a suggestion generation module. The project version matching module is used to match the project historical version according to the project basic information. The comparative analysis module is used to analyze project changes according to the project historical version. The suggestion generation module is used to generate change precautions and auxiliary suggestions according to project changes.

[0048] This embodiment also discloses a policy analysis method based on intelligent semantic recognition, which uses the policy analysis system based on intelligent semantic recognition of this embodiment.

[0049] Embodiment 2

[0050] The difference between this embodiment and embodiment one is that, in this embodiment, the suggestion generation module includes a change suggestion generation module, which determines the project change content, type and change trend according to the project change, generates a change reminder if the type is project requirement information and the project requirements become stricter, matches the corresponding suggestion model according to the specific change items, and generates a project change suggestion according to the suggestion model.

[0051] Embodiment 3

[0052] The difference between this embodiment and the second embodiment is that this embodiment also includes an enterprise database, which stores enterprise information of each enterprise, including basic enterprise information, such as name, address, identification number, etc., enterprise financial information, such as annual report, tax payment, audit report and other information, enterprise qualification information, such as enterprise honors, certificates, intellectual property rights, etc., enterprise personnel data, such as the number of personnel, educational distribution, title distribution, etc., enterprise historical data, such as historical financial data, historical personnel data, etc., the comparison and analysis module is also used to determine whether the enterprise meets the application conditions based on the enterprise information and project requirement information, and the suggestion generation module includes a cultivation suggestion generation module, and the cultivation suggestion generation module is also used to generate enterprise cultivation application suggestions and cultivation application plans for projects that do not meet the application conditions.

[0053] It also includes an enterprise development simulation module, which is used to simulate the development data of the enterprise in the next three years based on the enterprise information. The development data also includes enterprise financial information, enterprise personnel data, enterprise qualification information, etc. In this implementation, an artificial intelligence-based simulation model is used to simulate the development data. Specifically, it includes a BP neural network module, which is used to generate development data based on the enterprise's historical data. The BP neural network module includes a BP neural network model. The BP neural network model uses BP neural network technology to simulate the development of the enterprise. Specifically, a three-layer BP neural network model is first constructed, including an input layer, a hidden layer, and an output layer. In this embodiment, the basic information of the enterprise, the enterprise qualification information, the enterprise personnel data, the enterprise historical data, and the enterprise financial data are used as inputs of the input layer, and the output is the prediction of the corresponding data; this embodiment uses the following formula to determine the number of hidden layer nodes: Where l is the number of nodes in the hidden layer, n is the number of nodes in the input layer, m is the number of nodes in the output layer, and a is a number between 1 and 10. BP neural networks usually use Sigmoid differentiable functions and linear functions as the network's excitation functions. This paper selects the S-type tangent function tansig as the excitation function of the hidden layer neurons. The prediction model selects the S-type logarithmic function tansig as the excitation function of the output layer neurons. After the BP network model is constructed, the model is trained using the company's historical data as a sample. In order to improve the prediction accuracy, a classification prediction model can be established according to the company's field. The model obtained after training can obtain more accurate prediction results.

[0054] The comparative analysis module is also used to determine the type of unsatisfied application conditions based on the company's development data and unsatisfied application conditions. The types include unattainable, developable, and configurable. Unattainable means judging the conditions that the company cannot achieve within a certain time threshold based on the company's information. When such conditions occur, it means that the company cannot apply for the corresponding project, and there is no cultivation application suggestion or cultivation application plan. Developable conditions refer to conditions that can be met after a certain period of time, such as financial data, staffing, etc. As the company develops, it can meet the corresponding standards; Configurable conditions refer to conditions that the company can actively change and then achieve, such as lack of system, lack of certification system, etc. The company can meet the application conditions by optimizing the company system and seeking system certification.

[0055] The cultivation suggestion generation module generates cultivation application suggestions according to the application conditions of the developable type and the application conditions of the configurable type, and generates a cultivation application plan according to the development data simulated by the enterprise development simulation module.

[0056] Embodiment 4

[0057] The difference between this embodiment and the third embodiment is that it also includes a service agency database, which stores data of various service agencies. The cultivation suggestion generation module is also used to generate service cooperation recommendation suggestions. The cultivation suggestion generation module matches the corresponding service agencies according to the declaration conditions of the developable type and the declaration conditions of the configurable type, and generates a recommendation list. The cultivation suggestion generation module generates service cooperation recommendation suggestions based on the recommendation list. If the declaration condition of the configurable type is intellectual property, the corresponding agency is recommended; if the declaration condition of the configurable type is system certification, the corresponding certification service agency is recommended.

[0058] Embodiment 5

[0059] The difference between this embodiment and the fourth embodiment is that this embodiment further includes a complementary enterprise matching module and a complementary circle establishment module. The complementary enterprise matching module is used to match and group enterprises according to the information of each enterprise and the projects that the enterprises can apply for, so that enterprises with similar fields and complementary successful project experiences are assigned to the same group. For example, enterprise A has experience in successfully applying for project X and can now apply for project Y, while enterprise B has experience in successfully applying for project Y and now wants to apply for project X, so the two enterprises are assigned to one group. The complementary circle establishment module is used for enterprises in the group to establish a communication circle, so as to achieve mutual guidance on projects, reduce the project application costs of each enterprise, and improve the application success rate.

[0060] The above are only embodiments of the present invention. The invention is not limited to the fields involved in this implementation case. The common sense such as the known specific structures and characteristics in the scheme is not described in detail here. The ordinary technicians in the relevant field know all the common technical knowledge in the technical field to which the invention belongs before the application date or the priority date, can obtain all the existing technologies in the field, and have the ability to apply the conventional experimental means before that date. The ordinary technicians in the relevant field can improve and implement this scheme in combination with their own abilities under the enlightenment given by this application. Some typical known structures or known methods should not become obstacles for ordinary technicians in the relevant field to implement this application. It should be pointed out that for those skilled in the art, without departing from the structure of the present invention, several deformations and improvements can be made, which should also be regarded as the protection scope of the present invention, which will not affect the effect of the implementation of the present invention and the practicality of the patent. The protection scope required by this application shall be based on the content of its claims, and the specific implementation methods and other records in the specification can be used to explain the content of the claims.

Claims

1. Policy parsing system based on intelligent semantic recognition, characterized by: include: Policy acquisition module, used to obtain policy text; A project analysis module analyzes project information according to the policy text; the project information includes basic project information, project requirement information and project benefit information; The project parsing module includes a semantic recognition module, a splitting module and a parsing module; The semantic recognition module is used to perform semantic recognition on the policy text; The splitting module is used to split the policy text according to the semantic recognition result to obtain policy fragments of various types; The parsing module is used to parse the policy fragment according to the type of the policy fragment. The parsing module stores a project information identification data table. The project information identification data table stores a plurality of identification data. Each identification data corresponds to a type of project information. The parsing module obtains the corresponding identification data according to the type of the policy fragment. The identification data includes a parsing rule. The parsing module extracts the corresponding project information according to the parsing rule. Project version matching module, used to match project historical versions based on project basic information; An enterprise database storing enterprise information of each enterprise, including basic enterprise information, enterprise personnel data, and enterprise historical data; The enterprise development simulation module is used to simulate the enterprise development data in the next three years based on the enterprise information, and the development data includes enterprise financial information, enterprise personnel data, and enterprise qualification information; The comparative analysis module is used to analyze project changes based on historical project versions and determine whether the enterprise meets the application conditions based on enterprise information and project requirements; It is also used to determine the type of unsatisfied application conditions based on the enterprise's development data and unsatisfied application conditions, which include three types: unachievable, developable, and configurable; The suggestion generation module is used to generate change considerations and auxiliary suggestions based on project changes; The suggestion generation module includes a cultivation suggestion generation module; The cultivation suggestion generation module is used to generate cultivation application suggestions according to the application conditions of the developable type and the application conditions of the configurable type, and to generate a cultivation application plan according to the development data simulated by the enterprise development simulation module; A service agency database, wherein the service agency database stores data of various service agencies, wherein the cultivation suggestion generation module is further used to generate service cooperation recommendation suggestions, wherein the cultivation suggestion generation module matches corresponding service agencies according to the application conditions of the developable type and the application conditions of the configurable type, and generates a recommendation list, wherein the cultivation suggestion generation module generates service cooperation recommendation suggestions according to the recommendation list; A complementary enterprise matching module, which is used to match and group enterprises according to their information and the projects that they can apply for, so that enterprises with similar fields and complementary successful project experiences are assigned to the same group; A complementary circle establishment module is used for enterprises within the group to establish a communication circle.

2. The policy analysis system based on intelligent semantic recognition according to claim 1 is characterized by: The project information also includes other auxiliary information, and the project parsing module also includes other auxiliary information parsing modules. The other auxiliary information parsing modules include a key name parsing module and a key value parsing module. The key value parsing module extracts key values ​​according to preset parsing rules, and the key name parsing module obtains corresponding key names according to semantic recognition results and the position of key values ​​in policy texts.

3. The policy analysis system based on intelligent semantic recognition according to claim 2 is characterized by: The project requirement information includes time requirements, subject requirements and application method requirements. The subject requirements include field requirements, regional requirements, financial requirements, intellectual property requirements, establishment time, personnel requirements and qualification requirements.

4. The policy analysis system based on intelligent semantic recognition according to claim 3 is characterized by: It also includes a preprocessing module for performing formatting preprocessing on the acquired policy text data, wherein the formatting preprocessing includes removing stop words, removing duplicates, and correcting garbled characters.

5. The policy analysis system based on intelligent semantic recognition according to claim 4 is characterized by: It also includes a recognition model storage module and a model matching module. The recognition model storage module is used to generate and store a recognition model according to the data source of the policy text, the splitting result of the splitting module, and the parsing result of the parsing module; The model matching module is used to match the identification model according to the data source; The model matching module includes a model verification module, which is used to verify whether the recognition model is valid according to the semantic recognition result; The splitting module and the parsing module are further used to split and parse the policy text according to the effective recognition model based on the recognition module.

6. The policy analysis system based on intelligent semantic recognition according to claim 5 is characterized by: It also includes a project information identification data table expansion module, which is used to generate identification data and add it to the project identification data table based on the analysis results of other auxiliary information analysis modules.

7. The policy analysis system based on intelligent semantic recognition according to claim 6 is characterized by: The project information identification data table expansion module includes a parsing rule generation module, and the parsing rule generation module includes a basic rule generation module and an expansion module. The basic rule generation module is used to generate basic parsing rules according to key values ​​and key names, and the expansion module is used to expand the key values ​​and key names. The expansion module is also used to expand the basic parsing rules according to the expansion results of the key values ​​and key names to form parsing rules.

8. The policy analysis system based on intelligent semantic recognition according to claim 7 is characterized by: The project information identification data table expansion module also includes a standardization processing module, which standardizes each information key name according to a standardized vocabulary library.

9. A policy analysis method based on intelligent semantic recognition, characterized in that: A policy analysis system based on intelligent semantic recognition as described in claim 8 is used.

Images