A Machine Learning Method for Automatic Copyright Notice Identification of Source Files

Abstract

For Free and Open Source Software (FOSS), identifying the copyright notices is important. However, both the collaborative manner of FOSS project development and the large number of source files increase its difficulty. In this paper, we aim at automatically identifying the copyright notices in source files based on machine learning techniques. The evaluation experiment shows that our method outperforms FOSSology, the only existing method based on regular expression.