SysEvoRecomd: Graph Evolution and Change Learning Based System Evolution Recommender

In this paper, we present our approach System Evolution Recommender (SysEvoRecomd). In SysEvoRecomd, we used our proposed Graph Evolution and Change Learning (GECL) that does matrix reconstruction. Based on SysEvoRecomd, we developed an automation tool named as SysEvoRecomd-Tool, which is used to conduct experiments on various real-world evolving systems.

I. Introduction

An evolving system is a system which is continuously changing with time [1]–[3]. Constant growth of system make its maintenance a challenging task. Improper utilization of entities may lead to inefficient maintenance. Further, there is also a possibility of logical errors and faults in newer component due to many developers. The study of the system evolution can help to tackle these problems. An evolving system can be represented as a graph of entity-connections (a set of entities are vertices and directional connections among entities are edges).