Car Types and Semantics Classification Using Weka

Abstract

This paper presents a method of using Weka, a machine learning tool, to identify the difference of types and product semantics between fuel vehicles and electric vehicles. Pictures of 58 fuel vehicles and 42 pictures of electric vehicles during time period from 2021 to 2023 are selectively collected from Consumer Reports website to build the dataset. The fuel vehicle brands include Audi, BMW, Cadillac, and Lexus with 3 types, namely, SUV, Sedan and Luxury. In addition to the above-mentioned brands, Tesla is the 5th brand of electric vehicles. The perception of each picture is labelled by questionnaires of Automotive Model Semantics. Results reveal that even the smaller dataset can be trained to have highly accuracy models for classifying fuel vehicles and electric vehicles, different types as well as product semantics. The method presented is promising for studying car styling and exploring new applications of image classification to branding and product design.