Distributed investment of electric vehicles in Finland: A machine learning approach
School of Business | Master's thesis
Unless otherwise stated, all rights belong to the author. You may download, display and print this publication for Your own personal use. Commercial use is prohibited.
AbstractMethodologically, this paper presents a novel approach of employing Least Absolute Shrinkage and Selection Operator algorithm, combined with findings from the literature and data quality, first to eliminate regressors and only select the most relevant ones to later feed them into Logistic regression algorithm to construct a predictive model and predict the probability of adopting an electric vehicle (EV) for a random vehicle-owner. Empirically, using vehicle-level and individual-level administrative data for the year 2019 provided by Statistics Finland, this paper suggests that a vehicle-owner who has a tertiary education degree in the age bracket of [60-69] with (90th-100th] disposable income quantile living in urban municipalities is the most likely to adopt an EV. The least likely groups are demographically vehicle-owners who have a secondary education degree or below or is unknown and aged from 18 to 29 having disposable income within the (10th-20th] quantile living in rural municipalities and/or households who are a family unit or non-married individuals living alone in detached houses and/or individuals whose past and current vehicles were purchased in or before 2009. The gender of the vehicle-owner does not make a large difference in driving EV adoption behaviors. To reduce vehicle emissions, incentive-based and regulation-based environmental policies targeting these least likely demographic groups of vehicle-owner should be funded and prioritized by the Finnish government. JEL: [C51, C52, C53, D12]
Thesis advisorLiski, Matti
electric vehicle, machine learning, least absolute shrinkage and selection operator, LASSO, logistic regression, logit, predictive model, prediction