Construction method of chemical engineering knowledge graph

Abstract

The invention is suitable for the technical field of knowledge graph construction, and provides a chemical engineering knowledge graph construction method which comprises the following steps: collecting high-confidence data and converting the high-confidence data into a structured triple, and obtaining training corpora through a remote supervision training corpus construction method; extracting an e-book text by using a Python script, and extracting the e-book text as a data source through word segmentation and entity relationship extraction; extracting training corpora by using n-pattern mode features, limiting a type entity relationship extractor through a maximum entropy model, extracting a triple through the extractor, storing the triple into a Neo4j database, querying a matched entity relationship by using a Match clause in a Cyber language, and transmitting data to a front end in a JSON format through REST API; and the front end analyzes the data, and renders and interacts the knowledge graph by using an AntV G6 graph visualization engine. According to the method, data in the chemistry and chemical engineering field are integrated, the information query and analysis efficiency is improved, and discovery of new chemical knowledge is promoted.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of knowledge graph construction, and particularly relates to a construction method of a chemical engineering knowledge graph. BACKGROUND

[0002] A knowledge graph is a graphical data structure form for organizing, representing and storing knowledge, and its purpose is to enable computers to better recognize, understand and reason knowledge in imitation of the human cognitive and understanding way of knowledge, so as to present entities, relationships and attributes in the form of graphs. The technical bottom of the knowledge graph is a massive distributed database based on semantic web technology, with the content (i.e. semantics) of web data as the core, and linked in a way that machines can understand and process.

[0003] Due to the wide application of the knowledge graph, more and more fields use its related data to construct the required knowledge graph. The current chemical engineering knowledge graph is composed of multiple knowledge ontologies with chemical species (elements, compounds) as the core. The existing chemical knowledge graph is mainly constructed based on the knowledge of recording chemical species (molecular formula, charge, molecular weight and spin multiplicity, etc.) and the knowledge of representing chemical experiment information (various chemical experiment conditions), and has the following problems: the logical relationship between related entities is relatively simple, cannot represent more complex systems, reasoning ability and integration of new knowledge, and cannot solve chemical problems related to molecular engineering, chemical mechanism, multi-scale modeling, calculation and experimental automation, and interaction between chemists and machines. Therefore, the present application provides a construction method of a chemical engineering knowledge graph. SUMMARY

[0004] The present application aims to provide a construction method of a chemical engineering knowledge graph, which aims to solve the problems raised in the background.

[0005] To achieve the above-mentioned purpose, the present application provides the following technical solutions:

[0006] A construction method of a chemical engineering knowledge graph, comprising the following steps:

[0007] Step 1, data collection and processing;

[0008] Step 2, text data extraction and processing;

[0009] Step 3, feature extraction and model training;

[0010] Step 4, data storage and query;

[0011] Step 5, front-end display and interaction.

[0012] Further, the specific operation of step 1 is:

[0013] High-confidence data in the field of chemical engineering is collected, the high-confidence data is converted into structured triples, and a training corpus is obtained through a remote supervision training corpus construction method.

[0014] Further, the specific operation of step 2 is:

[0015] Text data of chemical engineering e-books is extracted through a Python script, then the text is subjected to natural language processing through a word segmentation tool, and then entity relationship extraction is performed, serving as a data source for establishing a knowledge graph.

[0016] Further, the specific operation of step 3 is:

[0017] n-pattern mode feature extraction is used to train the corpus, a maximum entropy model is used to limit the type of an entity relationship extractor, and the extractor is used to extract required triples from the encyclopedia text.

[0018] Further, the specific operation of step 4 is:

[0019] The extracted triples are stored in a Neo4j database, knowledge content is searched in the Neo4j database, a Match clause in a Cypher language is used to find entities and corresponding relationships consistent with the requirements, and then the data is transmitted to the front end in a JSON format through a REST API.

[0020] Further, the specific operation of step 5 is:

[0021] The front end parses the JSON data, uses a recursive algorithm to judge the hierarchical relationship of each node, and then uses an AntV G6 graph visualization engine to render and interact with the knowledge graph.

[0022] Compared with the prior art, the beneficial effects of the present application are:

[0023] The construction method of the chemical knowledge graph integrates data in the field of chemical engineering, improves the information query and analysis efficiency, and promotes the discovery of new chemical knowledge. BRIEF DESCRIPTION OF DRAWINGS

[0024] Figure 1 The flowchart of the present application. DETAILED DESCRIPTION

[0025] In order to make the purpose, technical scheme and advantages of the present application clearer, the present application is further described in detail below in combination with the drawings and examples. It should be understood that the specific examples described herein are only used to explain the present application and do not limit the present application.

[0026] The specific implementation of the present invention will be described in detail below with reference to specific embodiments.

[0027] like Figure 1 As shown, this invention provides a method for constructing a chemical engineering knowledge graph. This method adds (1) knowledge representing reaction mechanisms, such as information on reactants, products, and reaction processes; and (2) knowledge representing computational chemistry information, including descriptions of computational information such as the computational object, software used, computational method, and spin polarization. This integrates relevant chemical and chemical engineering data, improving information retrieval and analysis efficiency. It is expected to improve and enhance the properties of molecules or materials, build a complete ecosystem of chemical and chemical engineering knowledge, and effectively discover more new chemical knowledge through the exploration of the chemical knowledge space.

[0028] The construction method includes the following steps:

[0029] Step 1: Data Collection and Processing

[0030] Collect high-confidence data corresponding to relevant terms in literature, databases, and encyclopedias in the field of chemistry and chemical engineering;

[0031] Convert high-confidence data into structured triples, i.e. (entity, relation, entity), to facilitate subsequent processing and analysis.

[0032] By constructing a training corpus under remote supervision, initial relation triples are used to process the data through back-indexing, and then the required training corpus is obtained. This reduces human selection and ensures the objectivity and accuracy of the corpus.

[0033] Step 2, Text Data Extraction and Processing:

[0034] Extracting text data from e-books of chemical engineering and chemical engineering materials using Python scripts;

[0035] Then, natural language processing is performed on the text using word segmentation tools to break it down into words or vocabulary units;

[0036] Then, entity relationships are extracted and used as the data source for building the knowledge graph.

[0037] Step 3: Feature Extraction and Model Training

[0038] Extract training corpus using n-pattern features;

[0039] Further training yields maximum entropy model-bounded entity relation extractors, which are capable of extracting specific types of entity relations from text.

[0040] The extractor extracts the required triples from the encyclopedia text according to the desired relationship category, providing data support for the construction of geographic entity relationships.

[0041] Step 4: Data storage and retrieval:

[0042] The extracted triplet data is stored in the Neo4j database; this is a high-performance graph database suitable for storing network and complex relational data.

[0043] Search for knowledge content in the Neo4j database and use the Match clause in the Cypher language to find entities and corresponding relationships that match the requirements;

[0044] The data is then transmitted to the front end in JSON format via a REST API, allowing the data to be accessed and manipulated through a web application.

[0045] Step 5: Front-end display and interaction:

[0046] The front-end parses the JSON data and uses a recursive algorithm to determine the hierarchical relationship of each node, thus enabling the customization of graph icons, styles, categories, and nodes;

[0047] Then, the AntV G6 graph visualization engine is used to render and interact with the knowledge graph, allowing users to browse and explore knowledge and relationships in the field of chemistry and chemical engineering through a graphical interface.

[0048] The above are merely preferred embodiments of the present invention. It should be noted that those skilled in the art can make several modifications and improvements without departing from the concept of the present invention, and these should also be considered within the scope of protection of the present invention. These modifications and improvements will not affect the effectiveness of the implementation of the present invention or the practicality of the patent.

Claims

1. A method for constructing a chemical engineering knowledge graph, characterized in that, Includes the following steps: Step 1: Data collection and processing; Step 2: Text data extraction and processing; Step 3: Feature extraction and model training; Step 4: Data storage and retrieval; Step 5: Front-end display and interaction.

2. The method for constructing a chemical engineering knowledge graph according to claim 1, characterized in that, The specific operation of step 1 is as follows: We collect high-confidence data in the field of chemical engineering, convert the high-confidence data into structured triples, and obtain training data through a remote supervised training corpus construction method.

3. The method for constructing a chemical engineering knowledge graph according to claim 1, characterized in that, The specific operation of step 2 is as follows: Text data from e-books on chemistry and chemical engineering is extracted using Python scripts, then natural language processing is performed on the text using word segmentation tools, and finally entity relations are extracted to serve as the data source for building a knowledge graph.

4. The method for constructing a chemical engineering knowledge graph according to claim 2, characterized in that, The specific operation of step 3 is as follows: The training corpus is extracted using n-pattern features. A maximum entropy model is used to limit the type entity relation extractor, and the extractor extracts the required triples from the encyclopedia text.

5. The method for constructing a chemical engineering knowledge graph according to claim 4, characterized in that, The specific operation of step 4 is as follows: The extracted triple data is stored in the Neo4j database. Knowledge content is searched in the Neo4j database, and entities and corresponding relationships that match the requirements are found using the Match clause in the Cypher language. Then, the data is transmitted to the front end in JSON format via the REST API.

6. The method for constructing a chemical engineering knowledge graph according to claim 5, characterized in that, The specific operation of step 5 is as follows: The front-end parses the JSON data and uses a recursive algorithm to determine the hierarchical relationship of each node. Then, it uses the AntV G6 graph visualization engine to render and interact with the knowledge graph.

Images