Software project failures prediction using logistic regression modeling

The prediction of software project failure early can help in taking an enhancement steps that can steer the project outcome from failure to success. A range of risks may affect the software project during the development process and may lead to project failure. This paper presents a software project failure evaluation model developed based on real data collected from different software project reports, surveys and case studies. The constructed dataset describes the relationship between software project failure and independent failure factors. In this paper, the researchers have developed a failure prediction model using logistic regression method. This model can be used by project managers to assess the expected failures. The developed model helps in estimating the project outcome (Failed/Success). Furthermore, the model provides a probability of software project failure. The model is developed to enable the project decision makers to perform evaluation for the project status during any phase of the software development life cycle.