Enhancing software effort estimation through a hybrid approach of machine learning models and fuzzy logic
Year: 2026
Type: article
Abstract: Effective software project management requires accurate cost and effort estimation. An integrated strategy using fuzzy logic and machine learning improves software project cost and effort estimates in this paper. The proposed estimation method uses fuzzy logic and machine learning on Desharnais datasets. The result is obtained using an ensemble classifier that combines LR and LSTM models and fuzzy rule building. MAE, MSE, and RMSE are used to evaluate the suggested approach's efficacy in real-world software project planning. The LSTM model had the greatest MSE, RMSE, and MAE of 0.00806186, 0.08249961, and 0.0739960. In contrast, the LR model had MSE 0.00644958, RMSE 0.0528396, and MAE 0.0611382. The ensemble of LR and LSTM models enhanced accuracy and efficiency with a minimal MSE of 0.00281238, RMSE of 0.05971241, and MAE of 0.034620310. This hybrid method can change software project planning. This area of research has the potential to improve and expand the proposed approach, making it worth future study.
Authors: Ashish Kumar Jha, Rahul kumar
Institution: Ranchi University
Language: English
Subfield: Information Systems
Field: Computer Science
Domain: Physical Sciences
Open Access status: bronze
OpenAlex ID: W7214722628