Abstract: Cost estimation is a vital challenge in agile environments due to its dynamic nature. It has been noticed that existing agile approaches rely heavily on prior project data to estimate cost, size, effort, and duration. In prior work, we evaluated project-related and people-related elements to calculate project size and duration. If a software project is finished on time, under budget, and with the required level of quality, it will be deemed successful. When a project needs to change, cost estimation becomes more challenging. So, there is a need for an NFAH approach that may calculate the project's cost and work within budget and on time. To start using reliable project management software, it needs to take a whole different approach. In this work, we provide a Neuro-Fuzzy Agile hybrid, in which Artificial Neural Networks (ANNs) act as universal approximators to identify the non-linear function that will be used to train the datasets in the necessary format. Neuro-Fuzzy Agile Hybrid Model may find ambiguous inputs and give robustness in the face of imprecision because it combines neural networks, fuzzy logic, the Psycho-Fuzzy Inference System (PNFIS), Neuro-Fuzzy Bank (NFB), and the agile model. Accurate software estimating, including risk analysis, cost estimation, and quality assessment, is a major problem in software project development. According to the results, with an MMRE of 7.0745 and a Root Mean Square Error (RMSE) of 0.1191, the proposed model outperforms the Doty, Halstead, Walston-Felix, and Bailey-Basili models. The Neuro-fuzzy system-based model proposed in this study provides a more accurate estimation than the state-of-the-art models, which include the Neuro-fuzzy Agile Hybrid model (NFAH), Doty models, and software models from Halstead, Walston-Felix, and Bailey-Basili, respectively.

Keywords: Neuro Fuzzy Agile Hybrid Model, Halstead Model, Walston-Felix Model, Bailey-Basili Model, Doty Model, PNFIS, NFB, ANN

Pages: 497-511