A New Cost-Quality Estimation Model Based on Case-Based Reasoning Technique

Software cost, time and quality estimation is very important for developing new software projects. Accurate estimates of software cost, time and quality are required for developing software systems efficiently. A lot of estimation methods have been developed. Among those, COCOMO II, the model widely used for estimating effort and time, and COQUAMO, the model used to estimate the quality of the software project. These two models need to be calibrated when the dataset grows up to get new values for their parameters. Currently, neural network, and fuzzy logic modelling etc. are used for finding the accurate estimates. Since fuzzy logic approach is difficult to use and there are no clear guidelines for designing neural networks. In this paper, a new cost-quality model based on case based reasoning CBR technique is introduced for estimating software effort, time, and quality due to its dependence on historical information from past similar completed software project which means accurate estimates can be obtained. To estimate new software projects, the experiments have been conducted on NASA 93 dataset. As result, good estimates are obtained.