Data Mining and Knowledge Engineering

Sanjay Madan ¹, Purnima Ahuja ², Shalini Batra ³

Affiliations
1 Computer Science and Engineering, Thapar University, Patiala-147004, IN
2 Information Technology, Banasthali University, Banasthali, Rajasthan, IN
3 Computer Science and Engineering Department, Thapar University, Patiala-147004, IN

Abstract

To understand the software source code lots of approaches have been developed and many of them concern to the program structural information but this results in the loss of domain semantic crucial information contained in the text or symbols of source code. To understand software as a whole, we need to enrich these approaches with conceptual insights gained from the domain semantics. This paper proposes the mapping of domain to the code using the information retrieval techniques to use linguistic information, such as identifier names and comments in source code. Here we introduce the concept of Semantic Clustering, and an algorithm to group source artifacts based on how the synonymy and polysemy is related. Based on semantic similarity automatic labeling of the program code is done after detecting the clusters, and is visually explore in 3-Dimension format. The most important feature of theis approach is that it works at the source code textual level which makes it language independent. The approach correlates the semantics with structural information applies at different levels of abstraction (e.g.packages, classes, methods).

Keywords

Information Retrieval, Latent Semantic Indexing, Semantic Clustering, Software Reverse Engineering.

Full Text