A Feature Selection Based Relief Algorithm with Fuzzy Logic for Software Effort Estimation

Software effort estimation is a vital factor in any product industry. As programming gets developed in size and intricacy, it is extremely hard to precisely anticipate the cost of programming advancement. This was the difficulty in past years. The best entanglement of programming industry was the quick changing nature of programming advancement, which has made it hard to create parametric models that output high precision for programming improvement in all areas. This paper, proposes a novel technique to estimate software effort based on fuzzy logic (FL) along with relief algorithm. Relief algorithm is used to extract features. Mean Square Error and Accuracy are used as parameters to evaluate the results. Proposed technique is compared with various existing algorithms of software effort estimation and experimental results demonstrate that the proposed technique gives less error and hence provides better accuracy than the other existing techniques.