Extreme Gradient Booster Model (XGB) Model for CMMI Assessments

The Capability Maturity Model Integration (CMMI) model from the CMMI Institute, a subsidiary of ISACA, focuses on multiple levels of graded maturity. Level 4 and Level 5 are the highest levels of expected performance from the organization. The high maturity practices of the CMMI standards expect the use of statistical models for setting up, measuring, and corroborating organizational business performance. In the corporate world and the extant literature, organizations have mainly utilized regression-based Multilinear Regression models to set up and predict performance metrics as expected by the CMMI model, High Maturity Practices. This paper attempts to illustrate the effectiveness of the Extreme Gradient Boosting (XGB) model in precisely anticipating the Business Objectives required for CMMI Level 5 evaluations for a prominent software engineering company. The organization used complex engineering, operations, and business variables to create CMMI Business Objectives prediction models for CMMI Level 5 evaluations. The study showcases the effectiveness of the XGB model hitherto unused in the CMMI High Maturity practice in predicting the firm's performance on core business metrics, e.g., the revenue per employee and the gross margin.