Abstract:
"The North American railroad industry is vital for efficient goods transportation, acting as a backbone for the continent's economic activities. In response to rising demand, railroad companies must optimize operations for customer satisfaction and sustainable growth. Accurate forecasting of cargo transportation demands is pivotal, ensuring service quality, cost-effectiveness, and heightened customer satisfaction in this dynamic industry. The aim of this study was to find out a machine learning approach to predict volumes to be shipped on a particular month for a given origin and destination more accurately and efficiently by using a regression model and data analysis techniques to make the better marketing campaign for consumers, so that they can increase their business in areas with low predicted volumes.
This prediction model was developed with respect to the systematically collected data for past eight years on origin, destination, month, year, volume of a current databases by using complex SQL queries. A several regression models such as Linear Regression, SVM Regression, Decision Tree Regression, Random Forest Regression were adopted in training the data set for the above model. The Random Forest Regression was identified as the model with the best accuracy, with the accuracy score of 0.99. With the accuracy obtained, using the developed model it can be concluded that the ability of predicting the volume of an upcoming month using input data. Thereby, this model facilitates the customer by providing information on the railcars assigned for a particular month to fulfill and adjust their needs. On the other hand, it would be a valuable tool for the company to identify the areas of high demand and areas of low demand to tailor make the marketing strategies accordingly to boost the customer demand aiming future growth."
