Application of Ant Colony Optimization Techniques to Predict Software Cost Estimation
In modern society, machine learning techniques employed to predict Software Cost Estimation viz. Decision Tree, K-Nearest Neighbor, Support Vector Machine, Neural Networks, and Fuzzy Logic and so on. Every technique has contributed good work in the signif
- PDF / 175,325 Bytes
- 11 Pages / 439.37 x 666.142 pts Page_size
- 2 Downloads / 167 Views
Abstract In modern society, machine learning techniques employed to predict Software Cost Estimation viz. Decision Tree, K-Nearest Neighbor, Support Vector Machine, Neural Networks, and Fuzzy Logic and so on. Every technique has contributed good work in the significant field of software cost estimation. The Computational Intelligence techniques also contributed a great extent in standard-alone. Still there is an immense scope to apply optimization techniques. In this paper, we propose Ant colony optimization techniques to predict software cost estimation based on three datasets collected from literature. For each datasets, we performed tenfold cross validation on International Software Benchmarking Standards Group (ISBSG) dataset and threefold cross validation performed on IBM Data Processing Service (IBMDPS) and COCOMO 81 datasets. The method is validated with real datasets using Root Mean Square Error (RMSE).
Keywords Software Cost Estimation (SEC) Ant Colony Optimization Technique (ACOT) Travelling Sales Person (TSO) Root Mean Square Error (RMSE)
V. Venkataiah Computer Science and Engineering, CMR College of Engineering and Technology, Medchal, Hyderabad, India e-mail: [email protected] R. Mohanty (&) Computer Science and Engineering, Keshav Memorial Institute of Technology, Narayanaguda, Hyderabad, India e-mail: [email protected] J.S. Pahariya Computer Science and Engineering, Rustamji Institute of Technology, Tekanpur, Gwalior, India M. Nagaratna Computer Science and Engineering, JNTUH College of Engineering, Kukatpally, Hyderabad, India e-mail: [email protected] © Springer Nature Singapore Pte Ltd. 2017 S.C. Satapathy et al. (eds.), Computer Communication, Networking and Internet Security, Lecture Notes in Networks and Systems 5, DOI 10.1007/978-981-10-3226-4_32
315
316
V. Venkataiah et al.
1 Introduction Software cost estimation is an important tool which can impact the planning and budgeting of a project. Effective monitoring and controlling of a software project is required to estimate the cost, accuracy and quality. Accordingly, in modern society many machine learning techniques employed to find out the software cost estimation i.e. Decision Tree, K-Nearest Neighbor, Support Vector Machine, Neural Networks, and Fuzzy Logic and so on. Neural network [1–4] contributed good work over a decade in the significant field of software to predict cost, effort, and size estimation. These were trained and tested by Back propagation, gradient descent algorithm employed on different prominent datasets viz. ISBSG, IBMPDS, COCOMO 81, DESHARNAIS, CF, and so on. Followed by Fuzz Logic [2, 5–7] and rest of the techniques has also contributed. Further, the Computational Intelligence Standard-alone techniques [8] are Multiple Linear Regression (MLR), Polynomial Regression, Classification and Regression Tree (CART), Multivariate Adaptive Regression Splines (MARS), Radial Basis Function Neural Network (RBF), Counter Propagation Neural Network (CPNN), Dynamic Evolving Neuro– Fuzzy Inference System (DENFI
Data Loading...
