Identification of the Ideal Team Capabilities and Predictive Success Measure for Software Projects Using Machine Learning

Abstract

Software project management is always been a crucial aspect of modern software development research and practice due to the very high stake of time, financial constraints, and various external factors correlating to the outcome of the projects. In the recent past, multiple parallel research outcomes can be observed in terms of predicting the software project outcomes in the form of success or failure. Nonetheless, due to the wide variety of influencing parameters in the software project matrix, the prediction of the outcome is highly inaccurate. Thus, in this work, a novel multi-regression model is deployed to predict the outcome of the software projects. The proposed multi-regression method is a collaborative regression method, which deliberates the internal regression between the project matrix attributes and generates the most optimal regression model. The outcome of the proposed multi-regression method is nearly 98.3% and almost outperformed all the other parallel research outcomes. Also, this work delivers yet another outcome in form of a recommendation system for an optimal team design matrix based on various team capabilities. The recommendation system will analyze the existing human resources in the organization and based on the statistics of the project, client history, and human resource capabilities will produce the age-based diversity, experience-based diversity, and skill-based diversity for various project roles to make the software project management highly timely and reduce risk of project failures.