Software Defect Prediction: State of the Art Survey

Abstract

Software has evolved into a critical component in today's world. The quantity of faults in a software product is connected to its quality, which is also restricted by time and cost. In terms of both quality and cost, software faults are costly. The practice of tracing problematic components in software prior to the product's launch is known as software defect prediction. Defects are unavoidable, but we should strive to keep the number of defects to a bare minimum. Defect prediction results in shorter development times, lower costs, less rework, higher customer satisfaction, and more dependable software. As a result, defect prediction procedures are critical for achieving software quality and learning from prior errors.In this study, we conduct a review of the literature from the last two decades and look into recent advancements in the field of defect prediction.