A Semantic Approach for Mathematical Expression Retrieval

Abstract

Math search or mathematical expression retrieval has become a challenging task. Mathematical expressions are very complex, they are highly symbolic, and they have a semantic meaning that we should respect. In this paper, we propose a similarity search method for mathematical expression based on a multilevel representation of expressions and a multilevel search. We used the K-Nearest Neighbors with three types of distances to evaluate relevance between expressions. In the experimental level, the proposed system significantly outperforms statistical algorithms.