| Mathematical expressions are important carriers for technology information communication and vital components of scientific and technological literature.However,mathematical expressions are different from common texts,the current search technology and search engine for common texts cannot be used for searching scientific and technological documents featuring lots of mathematical expressions,thereby affecting the spread and communication of technology information and becoming one of key problems to be urgently solved in the current information search and artificial intelligence domains.Aiming at the preceding problem,dissertation studies and designs a model for searching scientific and technological documents based on mathematical expression features by analyzing document composition elements and implements the search using mathematical expressions as the query language.First of all,in the document search feature selection and extraction phase,extract important information in scientific and technological documents and organize it into a document information table and formula information table.Second,in the document index phase,use document feature information,formula feature information,and the mathematical formula parsing algorithm to obtain information about sub mathematical expressions and adopt the layered index thought to construct the document index table and sub expression index table.Then in the search phase,build feature vectors based on the feature information of mathematical query expressions and design document result sets found in different match modes such as accurate match,inclusion match,and similar match.Finally,based on the importance degree of query items in different places of documents and match degree,combine the tf-idf weight method and use the cosine similarity to calculate the similarity of scientific and technological documents to obtain orderly search results.The experiment result shows that the model realizes the goal of searching scientific and technological documents based on mathematical expression features. |
