A Comprehensive Software Project Cost Assessment Framework utilizing Global-Local Artificial Neural Networks Optimized with High-Level Target Navigation Pigeon-Inspired Optimization

Effective software project management depends on the prediction and mitigation of risk. These risks cannot be entirely eliminated, and a robust effort estimation model can significantly reduce the risks associated with task scheduling and resource allocation. We propose a Comprehensive Software Project Cost Assessment Framework that utilizes global local Neural Networks optimized using High-Level Target Navigation Pigeon-Inspired Optimization (SPCA-GLANN-HLTNPIO). The framework begins by collecting input data from the PROMISE data repository. The data were preprocessed using multi-aspect co-attentional Collaborative Filtering (MACACF) to effectively manage missing values. After preprocessing, the refined data are input into the global local artificial neural network (GLANN) to predict software project costs. To enhance prediction accuracy, High-Level Target Navigation Pigeon-Inspired Optimization (HLTNPIO) optimizes the GLANN parameters and improves the overall cost assessment. The performance of the SPCA-GLANN-HLTNPIO technique was evaluated using metrics, such as accuracy, precision, and mean absolute error. It achieves 26.36%, 20.69%, and 30.29% higher accuracy, along with 19.12%, 28.32%, and 27.84% higher precision, respectively, compared to established techniques such as SPE-GA-ANN, DD-ANN-SPRA, and SEE-AP-ANN. This framework represents a significant advancement in software project management, and provides a more accurate method for effort estimation and risk mitigation.