Regresar

A Machine Learning Model Comparison and Selection Framework for Software Defect Pbkp_rediction Using VIKOR

Abstract:

In today’s time, software quality assurance is the most essential and costly set of activities during software development in the information technology (IT) industries. Finding defects in system modules has always been one of the most relevant problems in software engineering, leading to increased costs and reduced confidence in the product, resulting in dissatisfaction with customer requirements. Therefore, to provide and deliver an efficient software product with as few defects as possible on time and of good quality, it is necessary to use machine learning techniques and models, such as supervised learning to accurately classify and pbkp_redict defects in each of the software development life cycle (SDLC) phases before delivering a software product to the customer. The main objective is to evaluate the performance of different machine learning models in software defect pbkp_rediction applied to 4 NASA datasets, such as CM1, JM1, KC1, and PC1, then de-terminate and select the best performing model using the MCDM: VIKOR multi-criteria decision-making method.