Based on Multi-dimensional Software Data and Complex Networks for Code Smells Detection

Code smell is the product of improper design and operation, which may be introduced in many situations. It will cause problems for further software development and maintenance. Currently, most code smells detection methods detect through a single type of software data. There are restrictions on detecting code smells with complex definitions and characteristics. In this paper, we propose a method of applying multi-dimensional software data. This method builds complex network through structural data and historical version data, and determines the code smell instances by searching the network. We designed two smells detection strategies and evaluated them in four open source projects. Judging from the results, we demonstrate that our approach has 23% and 15% higher F-measures on Shotgun Surgery and Parallel Inheritance Hierarchy than the existing most advanced detection methods. The detection method based on multi-dimensional software data and complex network is effective, and this software network can also provides guidance for data-driven software research.