A Neural Network-Based Model for Software Effort Estimation Using Data Mining Techniques

Abstract

Software effort estimation is a critical activity in software engineering, as inaccurate estimates often lead to cost overruns, schedule delays, and project failure. This paper proposes a software effort estimation model based on data mining techniques, specifically artificial neural networks (ANNs). The proposed model was developed and evaluated using benchmark datasets from COCOMO and Desharnais. Data preprocessing and supervised learning techniques were applied to train and test the neural network model.

The experimental results demonstrate that the proposed ANN-based model achieved a coefficient of determination (R²) of approximately 0.91, indicating a strong correlation between actual and estimated effort values.