抽象的な

Obtaining Optimal Software Effort Estimation Data Using Feature Subset Selection

Abirami.R, Sujithra.S, Sathishkumar.P, Geethanjali.N

Effort estimation is usually an odd job and it should be done with utmost care. It is basically carried out with large set of Software Effort Estimation data. But the problem with estimating such a large data set is very hard and there are many chances of errors. To avoid drawbacks in effort estimation, various techniques have been widely used. But no method produces absolute result for effort estimation. Since visualization is not possible in these cases of estimation, no method can be the best for estimating efforts efficiently. There are some problems like (a)some methods will be suitable for some kind of projects only and (b)some kind of methods will be suitable for smaller projects only (c)some methods can eliminate some necessary data and it leads to wrong estimations. It affects other processes like time and cost estimation, resource allocation, scheduling, etc. So the Software effort estimation data should be reduced using feature subset selection methods to make effort estimation in a better way. Feature selection is very important in various pattern classification problems. The feature selector is applied to select a subset of features from the large set of features. And also the selected subset should be sufficient to perform the estimation process. By using Ant Colony Optimization (ACO) method as a feature selector for obtaining optimized and reduced essential data, optimal solution can be produced for estimating the effort.