Information Retrieval for Texts with Formulas

Nowadays a large number of scientific articles, research texts and books have been converted into digital form and published on open-access web sites. Formulas are an important component of research works, especially in mathematics, in natural sciences and in engineering disciplines. But most of the well-known modern search engines do not implement mathematical retrieval of texts from formulas. A mathematical search engine could solve many problems for authors doing some research and also for educational purposes also. In this paper, an approach to creating a mathematical search system for collections of documents, including those containing scanned images, is proposed: software tools and libraries are selected that allow them to be automatically translated into a structured form with possible error corrections. The paper presents a method for constructing an index and a fuzzy search algorithm. The proposed approaches are implemented in a prototype software system.