Extended Angel: Knowledge-based Approach for Loc and Effort Estimation for Multimedia Projects in Medical Domain

Developing good multimedia applications is expensive, mostly in terms of time and degree of difficulty. Estimating the cost and effort for multimedia application development is a great challenge. One of the useful techniques for predicting effort and lines of code (LOC) required for developing multimedia projects is by using the technique of “estimation by analogies”. This technique involves characterizing the project for which an estimate is required. This characterization forms the basis for finding similar or analogous projects that have been completed before and for which effort and LOC values are known. The tool ANGEL uses analogy for cost and effort estimation. However the accuracy of estimates generated through this tool is poor under certain conditions. It is proposed to overcome the limitations of ANGEL and to improve the accuracy of estimation by using a knowledge based technique. In knowledgebased estimation, estimates are generated using techniques of both ANGEL and knowledge rules. To analyze the performance of the proposed knowledge based estimation tools, the source code of 20 multimedia projects in medical domain have been taken as test cases. For these projects various metric attribute values useful for cost and effort estimation are computed and the estimation is done using the tool ANGEL. Then the estimation is improved using the knowledge rules. The results of the knowledge based tool and ANGEL tool are compared to assess the performance improvement in the estimation process of the proposed technique.