Software Development Effort Estimation using SNS Algorithm
Diksha Jain, Pravali Manchala, Sarika Mustyala, Manjubala Bisi · 2024
Software development effort estimation (SDEE) is important in determining how much work will be required to finish a software development activity. However, the deployment of SDEE models has been hindered by the Curse of the Dimensionality problem. This problem leads to the inability of software researchers and practitioners to apply a precise effort classification benchmark to facilitate learning and understanding. In this paper, we propose a Social Network Search (SNS) based feature selection approach to estimate software development efforts to alleviate the curse of the dimensionality problem. We reduce the number of features using different variations of SNS on five different data sets. The performance of the proposed approach is compared with models which use all the features. It is found that SNS based feature selection has better performance than all features. Over different measures different variations of SNS algorithm gives better performance.