Modelling a machine learning classifier for predicting student’s entrepreneurial intentions

Abstract

Entrepreneurship study and research is an important aspect in booming economy like India because entrepreneurs provide necessary impetus and act as a catalyst in bringing economic growth. In this study we have designed a logistic regression based predictive classification algorithm to predict entrepreneurship affinity to start a new business. The Logistic Model uses a binary categorical variable i.e., interest towards starting a business in near future (Sure or Not Sure) as dependent variable and socio-economic factors, behavioural traits i.e., Risk bearing capacity, Creativity, Decision Making capacity, Leadership ability, Ease of Communication, Self-confidence, and Willingness to enter unfamiliar territory as independent variables. The Model also compares various feature selection methods in improving predictive accuracy. Data is collected from 321 students using a structured questionnaire and model predicts the significant factors that impact the entrepreneurship decision and probability of students having positive attitude towards starting a new business. Findings from the study revealed that the Gender of the respondents, attitude towards entrepreneurship and risk bearing capacity as the significant factors impacting the student’s intent towards starting a new business in near future.