Analysis Of Software Cost Estimation and Debt Management Based on Deep Learning Approaches
K. Ravikumar, K. Saravanakumar, Anand Viswanathan, Mathivanan Durai, S. Kiruthika Devi, S. Kalaiselvan · International Journal of Computational and Experimental Science and Engineering · 2025
The precision of Software Effort Estimates (SEE) is essential for planning, managing, and thoroughly assessing software projects, ensuring they stay on budget and schedule. Achieving accurate SEE results is vital for handling future software development tasks, addressing the challenges of overestimating and underestimating resources. The method employs machine learning (ML) assessment approaches that produce highly accurate results, evaluating based on metrics, data sets, and relevant attributes. This paper analyzes potential applications of data science in management accounting. With large amounts of data, deep learning techniques can overcome some of these limitations. Initially, we collected the dataset from a standard repository, and we started the first step of data preprocessing for reducing null and unbalanced values based on Mini-Maxi-score normalization (Mm-Z-score). The final stage is classification is based on SoftMax deep Scaling Gated Adversarial Neural Network (SmDSAN2) evaluating the cost estimation debt budget schedule and reducing the false rate for analyzing the It can predict the costs required to build or develop software cost based on SmDSAN2algorithm has shown high accuracy for Predict the necessary cost to develop Software cost and effort estimation. Estimation techniques are used by categorizing the estimation of projects created using the Fuzzy margin rate to identify various debts available in different organizations. The SmDSAN2 algorithm will help software companies follow the rules and standards and reduce the cost of the overall estimation.