DefinePK

Abstract

In the fast growing world understanding a situation is highly dynamic process. The decisions are to be made in an instantaneous fashion following different stages of risk and ambiguity. Project managers and leaders are forced to make such critical decisions that change the effect of the company. One such critical decision that involves rationalism and logical diagnosing is the process of choosing a feasible employee for a given project. The main objective of this research work is to propose strategies for optimally electing employees for an effective team formation and to understand the relationship between team’s success and given project’s success. The method adopted for achieving the central theme of the article is to develop a novel research model that wraps three dominant machine learning approaches to form a triangular hybridization for a better quality team formation. The three methods inspired are Artificial Neural Network (ANN), Decision Tree (DT), and a proposed method called Ensemble Decision Tree (EDT) which is a boosted decision tree using logit boost algorithm that are embedded into the proposed research model for achieving the desired goal. As a pilot scale attempt the model is validated by training and testing it over 474 freelancers from leading sites. The results infer that there exist a direct dependence of team and project success and the proposed EDT approach outperforms other two methods yielding an accuracy of about 87.34% in predicting the unknown sample as a valid or an invalid agent for the current project under consideration.