Machine Learning Model of Digital Transformation Index for Mexican Households

Abstract

Digital transformation refers to the change in all aspects of human society by the adoption of digital technologies. Different methodologies and measurements have been proposed to determine the level of digital transformation in regions or countries. In this work, we propose the creation of a digital transformation index for Mexican households using machine learning models for digital transformation measurement analysis and estimation. We include three dimensions in terms of the information and communication technologies infrastructure, availability of services, and usage. We also use a public dataset from the Mexican government to build and train three machine learning models. Experimental results validate that our methodology can deliver a digital transformation measurement using machine learning models consistently with $84\%$ of accuracy and $84\%$ of F1-score. We also prototype a simple web application using the best machine learning model found. We anticipate that measuring the digital transformation in companies, governments, and households allows better decisions in business intelligence and public policy.