Cognitive Estimation of Development Effort, Time, Errors, and the Defects of Software
Amit Kumar Jakhar, Kumar Rajnish · Walailak Journal of Science and Technology (WJST) · 2015
In the software industry, measuring the effort and time for developing of software is very challenging. Measuring development effort and time comprises several phases, but measuring the effort in each phase creates problems. It is also observed that estimation of the effort required for developing a project may be over-estimated or under-estimated. It can lead to enormous damage to the organization, with respect to budget and schedule. So, to address the aforementioned, a cognitive technique is proposed for measuring the development effort, time, and errors. After measuring the development effort, machine learning techniques: Bayesian Net, Logistic Regression, Multi-perceptron, SMO, and Lib-SVM are applied for software defect prediction. To estimate the software development effort and defects, NASA PROMISE: CM1, KC3, PC1, PC2, and JM1 datasets and devised datasets (proposed cognitive technique parameters of original datasets) are used. The experimental results of both the experiments prove the goodness of the proposed work of this paper.