An Empirical Evaluation of Assorted Risk Management Models and Frameworks in Software Development

2. Prominent Approaches For Risk And Defects Management In Software Development

To produce high quality software the prediction of defects is very important. Following techniques are used for prediction:

Machine Learning

Software risk management and defects prediction is very popular in software engineering as it helps to reduce the cost by predicting the defect at early phases. During the development of software for organizations defect prediction uses different Machine Learning Techniques (MLT). To make software reliable, the software should be defect free (Challagulla, Bastani, Yen, & Paul, 2005). The main goal of machine learning (ML) is to build up the algorithm of practical value and algorithms should be well-organized. These are used to build defect prediction model. In ML we not only deal with time and space but also with amount of data, so it is a data driven technique. Data plays a vital role here; data helps to determine that given set of data is used as training or testing purpose. Quality of software significantly improves using MLT in SDP. For improving quality, a large number of different flavours of machine techniques exist but no particular technique is better than other in term of SDP. Figure 1 shows a learning diagram.

Figure 1.

Learning diagram

Figure is depicting the Learning Diagram whereby the training and prediction process is presented effectually so that the further classification can be done with higher degree of performance and accuracy.