Predicting Motivation among GKP Recipients using Regression Techniques in Machine Learning: An Implementation of Rapid Miner

Abstract

Recently, predictive analytics has found a place in many research areas. From business to healthcare, the revolutions of predictive analytics studies are constantly evolving to help decision makers identify the problem and make a wise decision. While a bigger impact has been reported on the development of predictive models in business studies, there has been very little effort that investigates the deployment of predictive models by using machine learning approaches specifically involving SMEs. Small business entrepreneurs (SMEs) are among the entities most affected because of the COVID-19 pandemic. Hence, this study aims to develop a model that could predict the motivation score of GKP recipients based on three factors: satisfaction, perceived value, and perceived expectation. Four different regression models, namely Linear, Ridge, Lasso, and SVR, were developed and evaluated as the best model by using Rapid Miner machine learning software tools. The GKIPP grant dataset has been used as a case study for estimating the motive of the GKP grant to evaluate the outcomes of various regression models. The findings indicate that linear and SVR models have produced highly accurate predictions about the motivation of GKP recipients. They have also produced a high proportion of R-square scores across all regression models, which is highly encouraging.