Software Cost Estimation by a New Hybrid Model of Particle Swarm Optimization and K-Nearest Neighbor Algorithms

A successful software should be finalized by cost and time estimations. Software is a production which its approximate cost depends on expert workforces and professionals. The most important approximate software cost estimation (SCE) is related to the trained workforce. Creative nature of software projects and its abstract nature make extremely cost and time of projects difficult to be estimated. Various methods have been presented in the software project cost estimation for performing a software project in the area of software engineering. COCOMO II model is one of the most documented models among template-based methods that has been proposed by Bohm. Common methods for estimating the time and cost are essentially abstract, accordingly, providing new methods for SCE is required and necessary. In this paper, a new method is presented to solve the problem of SCE by using hybrid particle swarm optimization (PSO) algorithm and K-nearest neighbor (KNN) algorithm. The method was evaluated on 6 multiple datasets with 8 different evaluation criteria. Obtained results show the more accurate performance of the proposed method.