Assessing software quality through web comment search and analysis

Abstract

When reusing software resources appearing on the Internet, developers often encounter the problem that it is hard to know the quality of candidate software. In this case, developers usually want to search and find referable user comment on the Internet. To assist this process, we proposed a textual comment based software quality assessment approach in this paper. It could search and collect the user comments of the software resource on the Internet automatically. Furthermore, the sentiment polarity (positive or negative) of a comment is identified and all the comments are classified into positive or negative collection. Then the quality aspects which the comment talks about are extracted so as to draw out the merits and drawbacks of software resources. With these information, developers can do candidate software selection easier and quicker in the software repository. To evaluate our approach, we apply our approach on a group of open source software. The results show that our approach could achieve satisfying precision in software quality assessment. © 2013 Springer-Verlag Berlin Heidelberg.