Software Cost Estimation: A Comparative Study of COCOMO-II, Halstead and IVR Models
Iqtidar AliInstitute of Computer Sciences and Information Technology (ICS/IT), The University of Agriculture Peshawar, Pakistan.
Shah NawazInstitute of Computer Sciences and Information Technology (ICS/IT), The University of Agriculture Peshawar, Pakistan.
Mohib UllahInstitute of Computer Sciences and Information Technology (ICS/IT), The University of Agriculture Peshawar, Pakistan.
Rafiullah KhanInstitute of Computer Sciences and Information Technology (ICS/IT), The University of Agriculture Peshawar, Pakistan.
Muhammad TariqInstitute of Computer Sciences and Information Technology (ICS/IT), The University of Agriculture Peshawar, Pakistan.
Imran Ud DinInstitute of Computer Sciences and Information Technology (ICS/IT), The University of Agriculture Peshawar, Pakistan.
Arshad KhanInstitute of Computer Sciences and Information Technology (ICS/IT), The University of Agriculture Peshawar, Pakistan.
Gohar Rehman KhalilInstitute of Computer Sciences and Information Technology (ICS/IT), The University of Agriculture Peshawar, Pakistan.
Majid KhanInstitute of Computer Sciences and Information Technology (ICS/IT), The University of Agriculture Peshawar, Pakistan.
Corresponding Author:
Muhammad Tariq (tariqahmad825@gmail.com)
Abstract:
The comparison of the three best cost estimation models to be simulated for two prestigious organization datasets. The aim is to find out the best one in term of time and money. Multiple efforts in wastes grabs the attention of the researchers that a software project neither completed in time nor budget. Extra ordinary best models have been developed in the recent days. From online services to construction industry the field has spread and modified itself with the changes. Now the cost estimation industry has grown well. There are specific models for specific purposes. As the field has changed to Machine Learning (ML) and still it is not easy to find out the correct cost estimation of a project. This study has shown the comparison of COCOMO-II, Halstead and IVR cost models. NASA 93 and Turkish Industry datasets has been chosen. The evaluation performance has been checked through Magnitude Relative Error (MRE) and Mean Magnitude Relative Error (MMRE). From the simulation of these projects with these models, COCOMO-II performs outstandingly. So this study suggest that COCOMO-II is best enough for software cost estimation.
Keywords:
ML (Machine Learning); COCOMO (Constructive Cost Model); IVR (Interactive Voice Response); MRE (Mean Relative Error); MMRE (Mean Magnitude Relative Error)