ML-DRIVEN APPROACH FOR OPTIMIZING CUSTOMER PREDICTIONS IN THE AUTOMOBILE INDUSTRY

Abstract

In the increasingly competitive automobile industry, identifying and targeting the correct customer segments is critical for optimizing marketing strategies and resource allocation. This project presents a comprehensive Machine Learning and Regression framework designed to categorize customers into four distinct strategic groups (A, B, C, and D) based on demographic and socio-economic variables. Utilizing a dataset containing features such as Age, Profession, Graduation status, and Spending Score, the study employs a systematic data science pipeline encompassing rigorous data cleaning, exploratory data analysis (EDA), and advanced feature engineering. A key innovation in this study is the development of the Work_Experience_to_Age_Ratio, a derived feature that provides deeper insights into a customer's professional stability and purchasing power relative to their life stage. The methodology compares the performance of Logistic Regression (LR) and Random Forest Classification (RFC) models. Initial findings indicate that while LR provides a robust baseline for linear relationships, the RFC model excels at capturing the non-linear complexities inherent in consumer behavior. The results demonstrate that factors like Profession (notably Healthcare and Artistic sectors) and Spending Score are the primary drivers of segmentation. By automating the classification process, this approach enables automotive companies to move away from generic mass marketing toward hyper-personalized engagement. This data-driven model provides a scalable solution for predicting customer groups, ultimately leading to improved conversion rates, enhanced customer satisfaction, and a more efficient supply chain tailored to specific segment demands.