Design And Implementation Of Automatic Solution System For Geometric Problems Based On Knowledge MAP

In recent years,the research of knowledge map is more and more popular.With the application of knowledge map in all walks of life more and more widely,people pay more and more attention to the domain knowledge map.Among them,the knowledge map in the field of education will be an important direction of knowledge map research.The research goal of this thesis is to design and implement an automatic solution system based on the geometry problem of knowledge map.The main research contents are as follows:Construct conceptual knowledge map of elementary mathematics.Knowledge acqui-sition is to acquire knowledge in the field of primary mathematics.The source of knowl-edge in this thesis is mainly the current junior high school and senior high school mathe-matics teaching materials,while referring to the mainstream teaching auxiliary books on the market and the relevant information on the Internet.Then there is the extraction of entities and relationships,because the knowledge map is mainly for the following auto-matic solution system as the basis,and the mathematical solution needs to have accurate information.At present,the automatic or semi-automatic generation of knowledge map can not meet the accuracy requirements we need,so we use the manual way to extract the entities and relationships.Knowledge representation uses classes in Java language to represent entities and relationships.Each entity corresponds to an entity class,and each relationship corresponds to a relationship class.The knowledge storage uses the graph database neo4 j.A knowledge map conversion module is written to convert entity class and relationship class into node and relationship class respectively and store them in neo4 j.Knowledge visualization is realized by using the visualization function of neo4 j.An automatic solution system for geometric problems is designed.The system is based on knowledge map,there are three types of knowledge map,which are concept knowledge map,topic knowledge map and case knowledge map.The concept knowledge map has been introduced before.The topic knowledge map is a knowledge map generated by specific geometric topics,and the case knowledge map is a knowledge map generated by specific instances,similar to rules.The knowledge of these two knowledge maps comes from the description of the conditions and conclusions of topic and input respectively.The extraction of entity relationship is natural language processing.The knowledge rep-resentation is class object in Java language.Entity and relationship correspond to entity class and relation class respectively.The visualization functions of knowledge storage and knowledge visualization are neo4 j and neo4 j respectively.The way to solve the prob-lem is to put the case knowledge map into the case knowledge map database,make the case knowledge map and the case knowledge map in the database match and contrast with each other,the comparison is triple,check whether the case knowledge map contains all the condition triple groups in the case knowledge map,if so,the topic knowledge map will be expanded according to the conclusion triplet in the case knowledge map,the con-dition for the end of matching and comparison is that all the instance maps are matched,or the entities(relationships)needed to generate the topic are generated,or the time-out is required.There are 409 entities and 3398 relationships in the conceptual knowledge map,and the correct rate of problem solving is about 61.5 %.Basically complete the objectives set by the initial system.