Scientific and technological achievement conversion integrated information system based on block chain

Abstract

The invention relates to the technical field of big data, in particular to a block chain-based scientific and technological achievement conversion integrated information system, which comprises a crawling module, a conversion module, an extraction module, a reasoning module, an arrangement module, a comparison module, a calculation module and a pushing module. Converting the policy text data into structured data, constructing an industrial knowledge graph based on the industrial entity relationship, obtaining enterprise information of the to-be-recommended enterprise, reasoning enterprise associated information of the to-be-recommended enterprise from the industrial knowledge graph based on the enterprise information, and recommending the to-be-recommended enterprise to the enterprise. Policy data and enterprise data are sorted and compared, enterprise information and enterprise associated information are used as target enterprise data of an enterprise to be recommended, and then data with high similarity is transmitted to a calculation module for matching through the calculation module; and the matching process of the calculation module is shortened and the pushing efficiency is improved through a mode of sorting and comparing in advance.

Description

Technical Field

[0001] This invention relates to the field of big data technology, and in particular to an integrated information system for the transformation of scientific and technological achievements based on blockchain. Background Technology

[0002] To create a favorable business environment, government departments at all levels will introduce various policies to benefit enterprises in a timely manner. However, due to the large number of policies, difficulty in finding them, and difficulty in understanding the various clauses of the policies, the healthy development of enterprises is affected. Therefore, it is necessary to use knowledge graphs to accurately recommend policies to enterprises.

[0003] However, by adopting the above method, the existing push system directly performs subsequent operations such as tag generation after obtaining policy data and enterprise data, without sorting and comparing them, which affects the timeliness of data processing and results in low efficiency of push results. Summary of the Invention

[0004] The purpose of this invention is to provide a blockchain-based integrated information system for the transformation of scientific and technological achievements, which aims to solve the problem that existing push systems directly perform subsequent operations such as tag generation after obtaining policy data and enterprise data without sorting and comparing them, thus affecting the timeliness of data processing and resulting in low efficiency of push results.

[0005] To achieve the above objectives, the present invention provides a blockchain-based integrated information system for the transformation of scientific and technological achievements, including a crawling module, a transformation module, an extraction module, an inference module, an organization module, a comparison module, and a calculation module;

[0006] The crawling module, the conversion module, the extraction module, the reasoning module, the sorting module, the comparison module, and the calculation module are connected in sequence;

[0007] The crawling module is used to crawl policy information based on the corresponding website and then obtain the corresponding policy data set;

[0008] The conversion module is used to generate tags for the acquired policy data set, then obtain the corresponding policy category tags, and convert the policy tags of the corresponding categories into structured data to obtain the data corresponding to the structured policies;

[0009] The extraction module is used to receive industry documents, extract industry entities and their corresponding relationships from the industry documents, and construct a corresponding industry knowledge graph based on the industry entities and their relationships.

[0010] The reasoning module is used to obtain enterprise information of the enterprise to be recommended, reason the corresponding enterprise-related information from the industry knowledge graph based on the enterprise information, and use the enterprise information and enterprise-related information as the target enterprise data of the enterprise to be recommended.

[0011] The sorting module is used to sort and classify the acquired policy data and enterprise data according to the corresponding tag types and then manage them in a unified manner.

[0012] The comparison module is used to compare the similarity of similar types of data in the sorted policy data and enterprise data to obtain policy data and enterprise data with high similarity.

[0013] The calculation module is used to pair structural policy data and enterprise data with high similarity. When the matching degree reaches the corresponding preset threshold, the corresponding policy ontology data is sent to the enterprise to be recommended.

[0014] The conversion module includes a replacement submodule, a word segmentation submodule, a removal submodule, an output submodule, and a classification submodule, which are connected sequentially.

[0015] The replacement submodule is used to preset policy data word segmentation tools;

[0016] The word segmentation submodule is used to segment policy text data using a target word segmentation tool;

[0017] The removal submodule is used to remove stop words from the initial policy terms;

[0018] The input submodule is used to input the target policy terms into a pre-trained vector model;

[0019] The classification submodule is used to perform classification prediction based on the word vectors.

[0020] The extraction module includes an industry term acquisition submodule, an identification submodule, a conversion submodule, a fusion submodule, and a relationship determination submodule, which are connected sequentially.

[0021] The industry term acquisition submodule is used to perform corresponding word segmentation processing on industry documents;

[0022] The recognition submodule is used to perform entity recognition operations on industry terms;

[0023] The transformation submodule is used to transform industry entities into entity vectors;

[0024] The fusion submodule is used to fuse the entity vectors and position vectors of any two industry entities in the same sentence of an industry document;

[0025] The fusion submodule is used to fuse the entity vectors and position vectors of any two industry entities in the same sentence of an industry document.

[0026] The reasoning module includes a matching submodule, a target enterprise entity determination submodule, and a generation submodule, which are connected sequentially.

[0027] The matching submodule is used to identify enterprise entities from enterprise information;

[0028] The target enterprise entity determination submodule is used to determine its presence in the industry knowledge graph;

[0029] The generation submodule is used to determine enterprise information relationships based on the target industry entity and the target enterprise entity.

[0030] The sorting module includes a data acquisition submodule, a data tag recognition submodule, and a data classification submodule, which are connected in sequence.

[0031] The data acquisition submodule is used to acquire tagged policy data and industry knowledge graphs of enterprise data;

[0032] The data tag recognition submodule is used to identify industry knowledge graphs of policy data and enterprise data with tags;

[0033] The data classification submodule is used to classify, organize, and summarize the product knowledge graphs of the identified policy data and enterprise data.

[0034] The comparison module includes a data comparison submodule and an algorithm learning submodule, which are connected to each other.

[0035] The data comparison submodule is used to compare the similarity between the categorized and organized policy data and enterprise data;

[0036] The algorithm learning submodule is used to collect comparison data during the comparison process of the comparison submodule and then add it to the database to increase the subsequent calculation speed.

[0037] This invention discloses a blockchain-based integrated information system for the transformation of scientific and technological achievements. The crawling module crawls policy information from corresponding websites to obtain a corresponding policy data set. The transformation module generates tags on the obtained policy data set to obtain corresponding policy category tags. Based on these policy tags, it transforms them into structured data to obtain data corresponding to the structured policies. The extraction module receives industry documents, extracts industry entities and their corresponding relationships from the documents, and constructs a corresponding industry knowledge graph based on these relationships. The reasoning module obtains enterprise information of the companies to be recommended, infers relevant enterprise information from the industry knowledge graph based on this information, and uses the enterprise information and its associated information as the basis for recommending the companies. The system uses target enterprise data for industry. The sorting module is used to sort and classify the acquired policy data and enterprise data according to the corresponding tag types and manage them in a unified manner. The comparison module is used to compare the similarity of the sorted policy data and enterprise data to obtain policy data and enterprise data with high similarity. The calculation module is used to pair the structural policy data and enterprise data with high similarity. When the matching degree reaches the corresponding preset threshold, the corresponding policy ontology data is sent to the enterprise to be recommended. By combining and comparing in advance, the system increases the efficiency of push and solves the problem that the existing push system directly performs subsequent tag generation and other operations after acquiring policy data and enterprise data without sorting and comparison, which affects the timeliness of data processing and leads to low efficiency of push results. Attached Figure Description

[0038] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below.

[0039] Figure 1 This is a schematic diagram of the structure of the blockchain-based integrated information system for the transformation of scientific and technological achievements, as presented in this invention.

[0040] 1-Crawling module, 2-Conversion module, 3-Extraction module, 4-Reference module, 5-Organization module, 6-Comparison module, 7-Calculation module, 201-Replacement submodule, 202-Word segmentation submodule, 203-Removal submodule, 204-Output submodule, 205-Classification submodule, 301-Industry term acquisition submodule, 302-Identification submodule, 303-Conversion submodule, 304-Fusion submodule, 305-Relationship determination submodule, 401-Matching submodule, 402-Target enterprise entity determination submodule, 403-Generation submodule, 501-Data acquisition submodule, 502-Data tag recognition submodule, 503-Data classification submodule, 601-Data comparison submodule, 602-Algorithm learning submodule. Detailed Implementation

[0041] Please see Figure 1 ,in, Figure 1 This is a schematic diagram of the module structure of the blockchain-based integrated information system for the transformation of scientific and technological achievements.

[0042] The present invention discloses a blockchain-based integrated information system for the transformation of scientific and technological achievements, comprising a crawling module 1, a transformation module 2, an extraction module 3, an inference module 4, an organization module 5, a comparison module 6, and a calculation module 7. The aforementioned solution solves the problem that existing push systems directly perform subsequent operations such as tag generation after obtaining policy data and enterprise data without organizing and comparing them, thus affecting the timeliness of data processing and resulting in low efficiency of push results.

[0043] In this specific embodiment, the crawling module 1, the conversion module 2, the extraction module 3, the reasoning module 4, the sorting module 5, the comparison module 6, and the calculation module 7 are connected in sequence;

[0044] The crawling module 1 is used to crawl based on the policy information of the corresponding website and then obtain the corresponding policy data set;

[0045] The conversion module 2 is used to generate tags for the acquired policy data set, then obtain the corresponding policy category tags, and convert the policy tags of the corresponding categories into structured data to obtain the data corresponding to the structured policies;

[0046] The extraction module 3 is used to receive industry documents, extract industry entities and their corresponding relationships from the industry documents, and construct a corresponding industry knowledge graph based on the industry entities and their relationships.

[0047] The reasoning module 4 is used to obtain enterprise information of the enterprise to be recommended, reason the corresponding enterprise-related information from the industry knowledge graph based on the enterprise information, and use the enterprise information and enterprise-related information as the target enterprise data of the enterprise to be recommended.

[0048] The sorting module 5 is used to sort and classify the acquired policy data and enterprise data according to the tag type and then manage them in a unified manner.

[0049] The comparison module 6 is used to compare the similarity of corresponding similar types of data in the sorted policy data and enterprise data to obtain policy data and enterprise data with high similarity.

[0050] The calculation module 7 is used to pair structural policy data and enterprise data with high similarity. When the matching degree reaches the corresponding preset threshold, the corresponding policy ontology data is sent to the enterprise to be recommended.

[0051] In this embodiment, the crawling module 1 is used to crawl policy information based on the corresponding website to obtain the corresponding policy data set. The conversion module 2 is used to generate tags for the obtained policy data set to obtain the corresponding policy category tags. Based on the policy tags of the corresponding categories, it is converted into structured data to obtain the data corresponding to the structured policies. The extraction module 3 is used to receive industry documents, extract industry entities and their corresponding relationships from the industry documents, and construct the corresponding industry knowledge graph based on the relationships between industry entities. The reasoning module 4 is used to obtain the enterprise information of the enterprise to be recommended, reason the corresponding enterprise-related information from the industry knowledge graph based on the enterprise information, and use the enterprise information and enterprise association information as the target enterprise data of the enterprise to be recommended. The sorting module 5 is used to sort and classify the acquired policy data and enterprise data according to the corresponding tag types and manage them in a unified manner. The comparison module 6 is used to compare the similarity of the sorted policy data and enterprise data to obtain policy data and enterprise data with high similarity. The calculation module 7 is used to pair the structural policy data and enterprise data with high similarity. When the matching degree reaches the corresponding preset threshold, the corresponding policy ontology data is sent to the enterprise to be recommended. By combining and comparing in advance, the push efficiency is increased, which solves the problem that the existing push system directly performs subsequent tag generation and other operations after acquiring policy data and enterprise data without sorting and comparison, thus affecting the timeliness of data processing and resulting in low push efficiency.

[0052] The conversion module 2 includes a replacement submodule 201, a word segmentation submodule 202, a removal submodule 203, an output submodule 204, and a classification submodule 205, which are connected in sequence.

[0053] The replacement submodule 201 is used to preset a policy data word segmentation tool;

[0054] The word segmentation submodule 202 is used to perform word segmentation on policy text data using a target word segmentation tool;

[0055] The removal submodule 203 is used to remove stop words from the initial policy terms;

[0056] The input submodule is used to input the target policy terms into a pre-trained vector model;

[0057] The classification submodule 205 is used to perform classification prediction based on the word vectors;

[0058] In this embodiment, the replacement submodule 201 is used to preset a policy data word segmentation tool, and replace the corresponding word segmentation tool dictionary with a preset corresponding policy dictionary to obtain a word segmentation tool for the corresponding policy; the word segmentation submodule 202 is used to perform word segmentation on the policy text data through the target word segmentation tool to obtain multiple initial policy words; the removal submodule 203 is used to remove stop words from the initial policy words to obtain the corresponding target policy words; the input submodule is used to input the target policy words into a pre-trained vector model to obtain word vectors; and the classification submodule 205 is used to perform classification prediction based on the word vectors and output the corresponding policy category label.

[0059] Secondly, the extraction module 3 includes an industry term acquisition submodule 301, an identification submodule 302, a conversion submodule 303, a fusion submodule 304, and a relationship determination submodule 305, which are connected in sequence.

[0060] The industry term acquisition submodule 301 is used to perform corresponding word segmentation processing on industry documents;

[0061] The recognition submodule 302 is used to perform entity recognition operations on industry terms;

[0062] The conversion submodule 303 is used to convert industrial entities into entity vectors;

[0063] The fusion submodule 304 is used to fuse the entity vectors and position vectors of any two industry entities in the same sentence of an industry document;

[0064] The fusion submodule 304 is used to fuse the entity vectors and position vectors of any two industry entities in the same sentence of an industry document;

[0065] In this embodiment, the industry term acquisition submodule 301 is used to perform corresponding word segmentation processing on the industry document to obtain multiple corresponding industry terms; the recognition submodule 302 is used to perform entity recognition operation on the industry terms to obtain multiple corresponding industry entities; the conversion submodule 303 is used to convert industry entities into entity vectors and obtain the position information of industry entities in the industry document, and generate position vectors based on the position information; the fusion submodule 304 is used to fuse the entity vectors and position vectors of any two industry entities in the same sentence of the industry document to obtain fusion features; and the relationship determination submodule 305 is used to input the fusion features into a pre-trained convolutional neural network to obtain the output industry entity relationship.

[0066] Furthermore, the reasoning module 4 includes a matching submodule 401, a target enterprise entity determination submodule 402, and a generation submodule 403, which are connected in sequence.

[0067] The matching submodule 401 is used to identify enterprise entities from enterprise information;

[0068] The target enterprise entity determination submodule 402 is used to determine its presence in the industry knowledge graph;

[0069] The generation submodule 403 is used to determine enterprise information relationships based on the target industry entity and the target enterprise entity;

[0070] In this embodiment, the matching submodule 401 is used to identify enterprise entities from enterprise information, match enterprise entities with industry entities in the industry knowledge graph, and take the successfully matched industry entities as target industry entities; the target enterprise entity determination submodule 402 is used to determine industry entities in the industry knowledge graph that have a connection relationship with the target industry entity, and take them as target enterprise entities; the generation submodule 403 is used to determine the enterprise information relationship based on the target industry entity and the target enterprise entity, and generate enterprise association information according to the enterprise entity, the target enterprise entity, and the enterprise information relationship.

[0071] In addition, the sorting module 5 includes a data acquisition submodule 501, a data tag recognition submodule 502, and a data classification submodule 503, which are connected in sequence.

[0072] The data acquisition submodule 501 is used to acquire tagged policy data and industry knowledge graphs of enterprise data;

[0073] The data tag recognition submodule 502 is used to identify the industry knowledge graph of policy data and enterprise data with tags;

[0074] The data classification submodule 503 is used to classify, organize, and summarize the product knowledge graphs of the identified policy data and enterprise data accordingly.

[0075] In this embodiment, the data acquisition submodule 501 is used to acquire policy data with tags and industry knowledge graphs of enterprise data, the data tag identification submodule 502 is used to identify the industry knowledge graphs of policy data with tags and enterprise data, and the data classification submodule 503 is used to classify, organize and summarize the identified product knowledge graphs of policy data and enterprise data accordingly.

[0076] Finally, the comparison module 6 includes a data comparison submodule 601 and an algorithm learning submodule 602, which are connected.

[0077] The data comparison submodule 601 is used to compare the similarity between the classified and organized policy data and enterprise data;

[0078] The algorithm learning submodule 602 is used to collect comparison data during the comparison process of the comparison submodule and then add it to the database to increase the subsequent calculation speed;

[0079] In this embodiment, the data comparison submodule 601 is used to compare the similarity of the classified and organized policy data and enterprise data. The algorithm learning submodule 602 is used to collect the comparison data records during the comparison process of the comparison submodule to obtain training samples for computation and learning, and then increase its database to increase the subsequent calculation speed, so as to achieve the effect of quickly comparing and obtaining results.

[0080] The above-disclosed embodiments are merely one or more preferred embodiments of this application and should not be construed as limiting the scope of this application. Those skilled in the art can understand that all or part of the processes for implementing the above embodiments and equivalent changes made in accordance with the claims of this application still fall within the scope of this application.

Claims

1. A blockchain-based integrated information system for the transformation of scientific and technological achievements, characterized in that, It includes a crawling module, a conversion module, an extraction module, an inference module, an organization module, a comparison module, and a calculation module; The crawling module, the conversion module, the extraction module, the reasoning module, the sorting module, the comparison module, and the calculation module are connected in sequence; The crawling module is used to crawl policy information based on the corresponding website to obtain the corresponding policy data set; The conversion module is used to generate labels for the acquired policy data set, obtain corresponding policy category labels, and convert the policy labels of the corresponding categories into structured data to obtain the data corresponding to the structured policies. The extraction module is used to receive industry documents, extract industry entities and their corresponding relationships from the industry documents, and construct a corresponding industry knowledge graph based on the industry entities and their relationships. The reasoning module is used to obtain enterprise information of the enterprise to be recommended, reason the corresponding enterprise-related information from the industry knowledge graph based on the enterprise information, and use the enterprise information and enterprise-related information as the target enterprise data of the enterprise to be recommended. The sorting module is used to sort and classify the acquired policy data and enterprise data according to the corresponding tag types and then manage them in a unified manner. The comparison module is used to compare the similarity of similar types of data in the sorted policy data and enterprise data to obtain policy data and enterprise data with high similarity. The calculation module is used to pair structural policy data and enterprise data with high similarity. When the matching degree reaches the corresponding preset threshold, the corresponding policy ontology data is sent to the enterprise to be recommended.

2. The blockchain-based integrated information system for the transformation of scientific and technological achievements as described in claim 1, characterized in that, The conversion module includes a replacement submodule, a word segmentation submodule, a removal submodule, an output submodule, and a classification submodule, which are connected sequentially. The replacement submodule is used to preset policy data word segmentation tools; The word segmentation submodule is used to segment policy text data using a target word segmentation tool; The removal submodule is used to remove stop words from the initial policy terms; The input submodule is used to input the target policy terms into a pre-trained vector model; The classification submodule is used to perform classification prediction based on the word vectors.

3. The blockchain-based integrated information system for the transformation of scientific and technological achievements as described in claim 2, characterized in that, The extraction module includes an industry term acquisition submodule, an identification submodule, a conversion submodule, a fusion submodule, and a relationship determination submodule, which are connected sequentially. The industry term acquisition submodule is used to perform corresponding word segmentation processing on industry documents; The recognition submodule is used to perform entity recognition operations on industry terms; The transformation submodule is used to transform industry entities into entity vectors; The fusion submodule is used to fuse the entity vectors and position vectors of any two industry entities in the same sentence of an industry document; The fusion submodule is used to fuse the entity vectors and position vectors of any two industry entities in the same sentence of an industry document.

4. The blockchain-based integrated information system for the transformation of scientific and technological achievements as described in claim 3, characterized in that, The reasoning module includes a matching submodule, a target enterprise entity determination submodule, and a generation submodule, which are connected sequentially. The matching submodule is used to identify enterprise entities from enterprise information; The target enterprise entity determination submodule is used to determine its presence in the industry knowledge graph; The generation submodule is used to determine enterprise information relationships based on the target industry entity and the target enterprise entity.

5. The blockchain-based integrated information system for the transformation of scientific and technological achievements as described in claim 4, characterized in that, The sorting module includes a data acquisition submodule, a data label recognition submodule, and a data classification submodule, which are connected in sequence. The data acquisition submodule is used to acquire tagged policy data and industry knowledge graphs of enterprise data; The data tag recognition submodule is used to identify industry knowledge graphs of policy data and enterprise data with tags; The data classification submodule is used to classify, organize, and summarize the product knowledge graphs of the identified policy data and enterprise data.

6. The blockchain-based integrated information system for the transformation of scientific and technological achievements as described in claim 5, characterized in that, The comparison module includes a data comparison submodule and an algorithm learning submodule, and the data comparison submodule and the algorithm learning submodule are connected. The data comparison submodule is used to compare the similarity between the categorized and organized policy data and enterprise data; The algorithm learning submodule is used to collect comparison data during the comparison process of the comparison submodule and then add it to the database to increase the subsequent calculation speed.

Images