Research And Application Of Short Text Similarity Algorithm Based On Semantic Dependency Tree

The explosive growth of information resources in the Internet era is full of our lives.A large part of these information resources are textual information in the form of natural language,such as e-mail,web pages,e-books,etc.With the development of artificial intelligence,various techniques of natural language processing are applied to all aspects of the Internet,such as text data mining,search engines,etc.,and text similarity calculation is one of the core technologies of natural language processing.In this paper,the background and research status of natural language processing research are deeply understood and analyzed,and the related theories are summarized and studied.The research goal of this paper is established: short text similarity algorithm research.At the same time,combined with the online education platform project participated in the study,a subjective title automatic scoring algorithm based on short text similarity algorithm was proposed,and the algorithm was applied to the automatic scoring system of subjective questions for political public courses.The short text similarity algorithm proposed in this paper uses a semantic tree to represent a semantically complete short text.The semantic tree takes the core words in the short text as nodes and the semantic dependence between nodes as the weight between the tree nodes.According to the characteristics of the semantic tree,the similarity of the core words,the types of semantic dependencies,and the similarity of words with the same semantic dependence are considered to calculate the short text similarity.Based on the short text similarity algorithm,an auto-scoring algorithm for subjective questions is proposed.The algorithm divides the answer text according to certain characteristics,and then calculates the similarity by dividing the short texts one by one to obtain a scoring matrix.The matrix calculates the final score for the entire answer text.Finally,the short text similarity algorithm and the subjective automatic scoring algorithm proposed in this paper are applied to the automatic scoring system of subjective questions for collegepolitical public courses.This paper selects the answers and standard answers of the political test terminology and the teacher's truth.The score was used as experimental data to conduct experiments,and the experimental results were analyzed to verify the effectiveness of the algorithm.This paper aims to solve the problem of short text similarity calculation based on semantic level,and hopes to provide new ideas and new application forms for text processing technology in Chinese natural language processing.