Neural Network Based Approach for Software Defects Density Prediction

Bulbul Chaudhary


Faults in software systems perpetuate to be a major quandary. Unexpected software failures in anexecutable product are caused due to software defects. Kenning the causes of possible flaws as well as identifying common software development areas that may need heedfulness from the bootup of a project could preserve mazuma, time and work. In todays cutting edge struggle its obligatory to make sensible efforts to curb and knock down the number of defects in software engineering. The chance of initially estimating the thinkable faultiness of software could avail on orchestrating, supervising and executing software development activities [3]. To fixate on the defect density area, it is a critical business requisite which can avail find the defect density in software releases. These software defects may lead to deterioration of the quality which might be the underlying cause of failure. In this paper ANN toolbox has been proposed to quantify fault proneness and defect density. Two approaches i.e Feed Forward Artificial Neural Network and Cascade Forward Artificial Neural Network are used. The dataset used is elicited from PROMISE repository .This Data set consist of 498 projects with their 22 attributes. For validation we use RMSE(Root mean square error) as a performance function.

Full Text:



  • There are currently no refbacks.