Automatic Aspect-Based Sentiment Analysis for Motor Vehicle Sales Forecasting

Abstract

Abstract: For evaluating the vehicle's services and sales, the online review offers a bunch of information. The study looks at how a customer feels about the vehicle, which affects how many cars are sold. The brand image is damaged, and sales are affected when a review is wrong about a car, while positive reviews contribute to increased sales. The sales forecast process considers the online review data and previous sales data as potential sources for more precise sales predictions. Machine learning might speed up the time-consuming process of forecasting sales and understanding market trends. This method makes it easier to find the relevant words in postal correspondence relating to finance, automobile rules, environmental regulations, and customer service. In this study, we used a BERT (Bidirectional Encoder Representation Transformer) to collect consumer reviews of automobiles. The technique is made more visible and understandable by the ML algorithm, which also facilitates the creation and management of such assessments. So, the sales staff will have access to accurate information about the automotive industry, which will help them predict sales.

Index Terms: BERT; Sentiment Analysis; Machine Learning