The Predictive Models for Customer Classification based on Customer Personality

Personalized marketing is a crucial part of the marketing mix since it involves tailoring marketing strategies to each individual customer's unique demands to forge long-lasting relationships and stay competitive in the competitive economy. In this work, we present a prediction model for classifying chocolate consumers based on consumer personality data referred to theories: the NEO-PI-R (Big Five Model) and the human life values. Initially, the k-means algorithm is used for consumer segmentation into different groups based on consumer traits or personality, then classification algorithms, including decision trees, random forest trees, and deep learning in the RapidMiner tool, were adopted to build up the predictive models for classifying consumer groups. Our results show that, with an accuracy of 85.95%, the random forest tree provides the best classification performance. Additionally, as a guideline, we recommend that the chocolate business customize offerings to certain chocolate consumer groups to provide an effective marketing and branding strategy for their consumers.