A semi-automated filtering technique for software process tailoring using neural network

It is widely known that implementation of the software development process to fit a given environment is the key to develop software at the lowest cost and highest quality. In general, applying an off-the-shelf software development process or an organizational process to a specific project can cause a lot of overhead if no effort is made to customize the given generic processes. Even though the process tailoring activities are done before starting a project, they are not given high importance. These activities depend on several process engineers who have a lot of experience and knowledge about process tailoring. Because of this dependence on human experience, it takes a long time to have a tailored process fit the project. To decide whether a specific task should be part of a given project or not is very time-consuming. Therefore, we suggest a semi-automated process tailoring method, which uses the artificial-neural network-based learning theory to reduce this time. We have demonstrated the effectiveness of our process filtering technique with a case study using process tailoring historical data as learning data.