Performance comparison of Putnam model using new technology trends for software maintenance cost estimation. (3rd April 2022)

• Record Type: Journal Article
• Authors: Singh, Chamkaur; Kamini, ; Juneja, Neenu; Joshi, Prabal Kumar; Garg, Rachit
• Is Part Of: Journal of discrete mathematical sciences & cryptography. Volume 25:Number 3(2022)
• Page Start: 691
• Page End: 703
• Publication Date: 2022-04-03
• Subjects: 68W99; Cocomo -- Putnam model -- PSO -- PCA -- LDA -- Machine learning; Computer science -- Mathematics -- Periodicals; Cryptography -- Periodicals; Computer science -- Mathematics; Cryptography; Periodicals; 004.0151
• DOI: 10.1080/09720529.2021.2016220
• Languages: English
• ISSNs: 0972-0529

Abstract

Abstract: Now the concept of software estimation is added by concrete concepts of fuzzy logic and neural networks that help in monitoring as well as understanding the software development dynamic process. The goal of all these is to estimate the estimating software effort. Approximate maintenance and judgment costs for a project is defined by software maintenance cost estimation. Due to environmental, human, political and technical like number of involved variables for cost estimate calculation, the maintenance cost estimation will never be same. In this paper we have proposed a new approach using Putnam model which is an efficient software maintenance cost estimation model. For dataset acquisition we have considered Bug Id reports instances of Matlab to perform classification of Sentimental based on severity. After pre-processing the dataset Principal Component Analysis (PCA) has been used to compute and study the Eigenvector for feature extraction. Then instance selection is performed using Particle Swarm Opitmization. In order to reduce matrix size there is need to deal with instances selection that will make processing easy to deal with further proceeding input. Once an instance is selected the next step is to classify it. In this work we have used Linear Discrement Analysis (LDA) as classifier which is an efficient statistical method used in machine learning. For testing the proposed approach we have compared the results obtained using COCOMO and PUTNAM model that shows the use of Putnam model will improve the system in terms of various parameters.