Case Study of Model Selection on Customer Information Task Based on Machine Learning Algorithms

Abstract

In today's era of high-speed development, more and more companies are finding customers with different needs in the market. Due to the large market size, each company cannot tailor its market for each customer, it’s difficult for them to predict the customer’s need, so market segmentation has emerged. This paper will use a case study about an automotive company to choose a best prediction modelusing the information of existing products and customers. The company divides the customer into 4 segmentations A, B, C and D. The quantitative method of study will be used to find the relationship between products and customers. Logistic Regression,KNN,SVM, Random Forest, and Decision Tree are used to compute the accurate rate. Decision Tree model was found to be the most accurate and the accuracy is 53%. In this paper, business objectives were defined, features and distribution of data were explored, data were processed, relevant features were selected, data were modeled, and accurate values between five different models were calculated. These steps can help the company find the nearest algorithmic model that allows it to use the best marketing strategy for its customers.