Private Vehicle Ownership Prediction Using Regression and Time Series Analysis

Abstract

In this study, the number of private vehicle in Bangladesh from 2000 to 2018 was analyzed. As the traffic congestion and road accident have become the burning issues of nowadays developing countries like Bangladesh, it is found that the increasing number of private vehicle is one of the major reasons for all of these problems. So here, three prediction models regarding the number of the private vehicle were developed for the upcoming years. The data were analyzed through regression (linear, log, and log–log) and time series analysis (autoregression and moving average model) using the GDP and population as explanatory variables since they are considered as the major factors of increasing the private vehicle of a developing country. These models can become the indispensable factor for the policymakers, while they will construct transportation and urban management policies for reducing the problems. If the number of private vehicle for the upcoming years can be assumed, future policies regarding these problems will be more efficient. Among the five models, the log–log regression model produced less average error which is only 11.21%. In this model, the R-squared and adjusted R-squared values have been found 0.9271 and 0.9180, respectively. Using this model, the numbers of private vehicles can easily be assumed, and the trend of increasing the private vehicles in the upcoming years can also be observed. As the average error is only 11.21%, the difference between the predicted value and the real value can easily be neglected.