A Synergistic Approach to Software Effort Estimation Using Fuzzy Logic and Machine Learning

Abstract

Precise software effort estimation is crucial for efficient project management, impacting planning, budgeting, and resource allocation. Conventional models frequently need help dealing with software development's inherent unpredictability and intricacy. This study introduces a new hybrid model that combines the clarity of fuzzy logic with the predictive capabilities of machine learning to improve software work estimation. By employing the Takagi-Sugeno-Kang (TSK) fuzzy logic method, we can account for the vague and uncertain elements of effort estimation. At the same time, machine learning models handle the non-linear connections and interactions among project factors. Our approach consists of gathering and preparing past project data, creating distinct fuzzy logic and machine learning models, and then combining these models into a unified hybrid system. The hybrid model is compared to independent fuzzy logic and machine learning models, showcasing its superior performance in terms of Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), and R-squared metrics. Moreover, the hybrid model offers data that may be easily understood, assisting project managers in comprehending the fundamental aspects that affect effort estimations. The actual use of our methodology in real-life project settings reveals its capacity to decrease project overruns and enhance budgeting precision. This research enhances the profession by providing a reliable, precise, easy-to-understand tool for estimating software work, improving project management processes