Empirical Evaluation of machine learning techniques for software effort estimation
Neha Saini, Bushra Khalid · IOSR Journal of Computer Engineering · 2014
Accurate estimation of software development effort is a very difficult job.Both under estimation as well as over estimation can lead to serious consequences.So its very important to find a technique which can yield accurate results for software effort estimation.Here in our paper we have evaluated various machine learning techniques for software effort estimation like bagging, decision trees, decision tables, multilayer perceptron and RBF networks.Two different datasets i.e. heiatheiat dataset and miyazaki94 dataset have been used in our research.Decision trees outperform all other models in term of MMRE value.Results of machine learning algorithms can be different from dataset to dataset.