An Extraction Method Of Latex-based Mathematical Formulas On Web

With the continuous development of Internet technology, a large amount of information has appeared on the Internet. The search engines provide a platform for information retrieval and sharing, but the traditional search engines have been unable to meet the people’s growing need in some aspects, In the field of mathematics, the influence of math Web forum, Wiki and other social sites is growing in mathematics education, and there are lots of mathematical formulas in these websites. But the text search engine can not meet the requirements that people search mathematical formulas. Therefore, how to search the web mathematical formula is very important for learning and research in the mathematical aspects. The mathematical formula extraction is the premise and foundation for index system and it is an important step of the search engine.In these websites the forms of mathematical formulas mainly include LaTex, picture and so on. And this paper studies the extraction of LaTex-based mathematical formulas on Web. On the one hand,combining with the BNF expression, this paper puts forward the method of analyzing the characteristics of the extract containing LaTex formula automatically and finding out the characteristics of the web page contains LaTex formulas. On the other hand, according to the characteristics the formulas contains, this paper proposes a method of extracting and filtering LaTex mathematical formulas.This method can filter out the non mathematical formulas of the LaTex that exist in the extract content, and increase the accuracy of the formula of extraction. Through the experiment discovered, it found that the recall rate reaches75%and the precision rate comes to99%using this method.