Using Computing Intelligence Techniques to Estimate Software Effort

Abstract

In the IT industry, precisely estimate the effort ofeach software project the development cost andscheduleare count for much to the software company. Soprecisely estimation of man power seems to begettingmore important. In the past time, the ITcompanies estimate the work effort of man power byhumanexperts, using statistics method. However, theoutcomes are always unsatisfying the managementlevel.Recently it becomes an interesting topic ifcomputing intelligence techniques can do better in thisfield. Thisresearch uses some computing intelligencetechniques, such as Pearson product-momentcorrelationcoefficient method and one-way ANOVA methodto select key factors, and K-Means clustering algorithmtodo project clustering, to estimate the softwareproject effort. The experimental result show thatusingcomputing intelligence techniques to estimate thesoftware project effort can get more precise andmoreeffective estimation than using traditional humanexperts did.