Implementation and evaluation of optimized algorithm for software architectures analysis through unsupervised learning (clustering)
Qadeem Khan, Usman Akram, Wasi Haider Butt, Saad Ur Rehman · 2016
Software projects are getting more complex and thus it is very difficult for the companies to develop their projects alone. There are multiple heterogeneous systems which are different by multiple perspectives such as different users functionalities delivered different CASE tools, technology adopted for software development, and different platform used for deployment. Thus heterogeneous systems are very difficult to be analyzed and maintained. In this research, we are going to perform the architecture analysis through data mining approach. There are many algorithms used in data mining for analysis of data such as: Association, Sequential Pattern, Similar Sequences, Demographics Clustering, Neural Networks, and Decision Tree. In this research, we optimized these algorithms through proposing our own algorithm for clustering software architectures. Since the data set of software architecture is not available so hence we designed our own data set for this research.