Customer Classification in Indian Retail Sector-A Comparative Analysis of Various Machine Learning Approaches

摘要

The retail industry across the world is realizing that delivering high levels of service quality and achieving customer satisfaction is the key for a sustainable competitive advantage. Researchers have found positive relations between retail service quality dimensions and customer satisfaction. Identifying and classifying the retail customers as 'satisfied' or 'dissatisfied' according to the retail service quality dimensions would be useful to retailers in enabling strategic decision making in a competitive and dynamic environment. Retailers generate and collect a huge amount of customer data on daily transactions, customer-shopping history, goods transportation, consumption patterns, and service records in a relatively short period. The explosive growth of data requires a more efficient way to extract useful knowledge which can help the retailers to make better business decisions and to target customers who might be profitable to them. The concept of data mining has emerged as an effective technique for exploring large amounts of data to discover meaningful patterns and rules in various fields including retail. In this paper, the retail customers are classified into either 'satisfied' or 'dissatisfied' classes according to the retail service quality dimensions. The research presents a comparative study of popular classification techniques such as decision tree classifier and support vector machine using the R-studio software. The paper uses machine learning algorithms to assess the Indian retail service quality. The results would help the retail organizations to enhance their overall service quality and to target their marketing efforts at the right group of customers.