Prediction of Software Effort in the Early Stage of Software Development: A Hybrid Model

The key challenge that project managers face during software development is the accurate prediction of the software effort. Improper prediction leads either to overestimation or underestimation of the software effort, which can have disastrous consequences for the stakeholders. This article attempts to design a model that gives an accurate prediction of effort in the initial phase of the software development lifecycle. The proposed model uses multilayer perceptron (MLP) and the generalized linear model (GLM) with the ensemble technique for the learning purpose. The model is trained and validated using the ISBSG dataset. The proposed model is compared for performance with two baseline models: MLP and GLM. The results show that the proposed model outperforms most of the baseline models against different performance metrics.