Method for mathematical expression retrieval based on hierarchical indexing

Abstract

The invention provides a method for mathematical expression retrieval based on hierarchical indexing. The method is realized based on indexes which are established in advance. Establishment of the indexes comprises that a mathematical expression of a LaTeX format is analyzed, a Hash value hierarchical structure tree of the expression is formed, and a Treap tree is constructed by inserting knots to form a KP mapping index layer and a reverse index layer. During the retrieval, the mathematical expression of the LaTeX format input by a user is analyzed in the same manner, the Hash value hierarchical structure tree is formed, and then the retrieval is conducted in the KP mapping index layer and the reverse index layer of the indexes so as to realize the accurate retrieval or the isostructural retrieval of the mathematical expression. The method for the mathematical expression retrieval based on the hierarchical indexing provided by the invention adopts a double-layer indexing structure, gives consideration to information integrity of the mathematical expression and efficiency and functions of the retrieval, has abundant modes of the retrieval and obtains high retrieval efficiency. The method for the mathematical expression retrieval based on the hierarchical indexing provided by the invention is funded by the Natural Science Foundation of China (Approval No. of Project: 61375075).

Description

technical field [0001] The invention relates to the field of mathematical information retrieval, in particular to a hierarchical index-based mathematical expression retrieval method. Background technique [0002] At present, some research institutions or organizations at home and abroad have begun to conduct related research on mathematical information retrieval, and have built prototype systems with mathematical information retrieval functions, such as DLMF Search, EgoMath, MathDex, LeActiveMath, MathWebSearch, WikiMirs, etc. These prototype systems can be divided into two categories according to the research and development strategies of mathematical expression retrieval technology: one is based on the full-text search engine with mathematical expansion, and the other is specially designed for the characteristics of mathematical expressions. Among them, DLMF Search, EgoMath, MathDex and LeActiveMath belong to the former category, and MathWebSearch and WikiMirs belong to th...

AI Technical Summary

Problems solved by technology

The mathematical expression retrieval method specially designed for the characteristics of mathematical expressions analyzes and transforms mathematical expressions by constructing a mathematical expression semantic structure tree, and builds an index for the purpose of retrieving mathematical expressions with similar semantics; its retrieval ability Depending on the quality of mathematical expression parsing and transformation and the performance of the specially constructed mathematical index and matching model, some transformations made to extract mathematical semantic information during index construction may cause the loss of the original hierarchical structure information of the expression, or mathematical expression hierarchy effects on the order of operation of expression operators are ignored

Images