Supervised and unsupervised learning applied to crowdfunding

Abstract

This paper aims to establish the participation behavior of residents in the city of Bogotá between 25 and 44 years of age, to finance or seek funding for entrepreneurial projects through crowdfunding? In order to meet the proposed objective, the focus of this research is quantitative, non-experimental and transactional (2017). Through data collection and data analysis, we seek patterns of behavior of the target population. Two machine learning techniques will be used for the analysis: supervised learning (using the learning algorithm of the decision tree) and unsupervised learning (clustering). Among the main findings are that (i) most of the people who would participate as an entrepreneur and donor and entrepreneur simultaneously belong to stratum 3; (ii) Crowdfunding projects based on donations do not have a high interest on the part of Bogotans, but those in which they aspire to recover the investment.