A research on the text comparison method using cosine similarity

Abstract

Currently, there are a number of solutions for detection of copy and a few software tools that allow detecting whether a document (called writing checks) has copied from a set of source materials or not. These methods are mainly based on the string search and matching, only really effective if the copy is “verbatim”. This paper proposes the method of text similarity comparison with cosine measure, an effective method to detect copying document, contributing to limit the infringement of intellectual property rights on scientific publications at present.