Implementation of using classification Data Mining Techniques for Software Cost Estimation

Abstract

Cost estimation is one of the biggest problems faced by software industry. It is necessary to have accurate cost estimation in order to conduct well budget. Under-estimation may lead to unexpected increase in budget, delay of project completion or its low quality, while over-estimation may lead to losing business opportunities. In this dissertation the idea of building data mining techniques into existing software cost estimation model such as COCOMO II model is implemented. Data mining allows users to analyze data from many different dimensions or angles, categorize it, and summarize the relationships identified.