Software Defect Prediction Using Gain Information Weighted

Software defect prediction activity concerns metrics of successful of test phase in Software Development Life Cycle (SDLC), companies work on predict the number of defects in software systems, there are some issues of predicting software defects such as imbalanced dataset which contains noisy attributes or lack information about all attributes, the noisy leads to decrease significantly results, This paper proposed a model designed to reduce the number of features used in the prediction process. The model will apply the method of gain information weighted to solve pervious issues reduce the number of selection features using the Gain information weighted, prediction using most popular algorithm of machine learning, prediction using random forest algorithm reaches 88%.