Design And Implementation Of Scholar Information Aggregation Query Platform Based On Knowledge Graph

The knowledge graph aims to describe various entities or concepts and their relationships in the real world,forming a huge semantic network graph.This thesis is oriented to the field of university scholars.Aiming at the problems of inconsistent sources of unstructured information in the field of scholars,various and messy forms of information,resulting in fuzzy relations between scholars,and insufficient orientation in obtaining scholar information,the scholar information is uniformly modeled and the implications are explored.The relationship between the scholar’s knowledge graph is formed,and an intelligent query algorithm is proposed based on the constructed knowledge graph,and then the scholar’s knowledge graph and query algorithm are integrated and researched to build a scholar’s information aggregation query platform to enhance the directionality of information acquisition and improve The accuracy of traditional query methods meets the diversification of query methods and has a certain degree of fault tolerance.It is convenient for enterprises,universities or other users to understand scholar information in a targeted manner,and quickly find scholars that meet their own R&D needs for auxiliary guidance.Enterprises,universities and other production applications have made meaningful research and exploration.The work done in this thesis is mainly divided into the following three parts:(1)The design and storage of scholars’ knowledge graph.By analyzing the acquired text data and defining entity attributes and their relationships according to their own needs,the pattern structure of the scholar’s knowledge graph is determined.Considering that some automated knowledge extraction methods are not sufficiently targeted in scholars’ text information,and the extraction effect is not very satisfactory,so combined with the method of manually defining rules,summarizing the rules according to the text content,constructing the extraction of entity attributes and relationships Rules,for rule-based attribute and relationship extraction.Aiming at the problems of poor performance in the storage and retrieval of scholar entity relationships when using relational databases for data storage,and low efficiency in database maintenance,this thesis uses graph database Neo4 j to store the acquired scholar knowledge and reduces the use of association tables.,To construct a knowledge graph of scholars.The experiment verifies the accuracy of the rule-based knowledge extraction method in this thesis,and the knowledge graph constructed is feasible.(2)Design of intelligent query algorithm based on natural language questions.The query algorithm uses semantic analysis of the query questions to accurately identify the query intent and improve the deviation of query results caused by improper application of search terms.Intent identification uses question classification and matching question templates,and defines six types of question templates and corresponding six types of Cypher query sentence generation templates based on the actual situation,and presets common question expression sentences and their characteristic words as training samples.In view of the long timeconsuming training of the classifier and low classification efficiency,the Spark MLlib machine learning component combined with the naive Bayes classification algorithm is introduced to train the problem classification model.Use keywords and classification labels obtained after semantic analysis and classification of query questions to generate Cypher query statements,perform queries,and return results.A comparative experiment was performed to verify the accuracy of the problem classification model mentioned in this thesis and the time-consuming performance of training.The experimental results show that when the data size is large,the Spark-based naive Bayes classifier can improve the training speed and classification accuracy.Thereby improving the accuracy and credibility of the query results,and providing an algorithm basis for the following query platform construction.(3)Design and implementation of scholar information aggregation query platform.Based on the scholar’s knowledge graph constructed above and the designed intelligent query algorithm,through demand analysis and framework design,the Springboot framework is used to build a query platform that separates front and back ends.The platform can realize the visual display of query results and entity relationship topological graphs through precise query of specific entity names or fuzzy query based on natural language questions.The function of the platform is verified through experiments,and the feasibility and practicability of the query platform are tested at the same time.