Correlation-augmented Naïve Bayes (CAN) Algorithm: A Novel Bayesian Method Adjusted for Direct Marketing

• Published in: Applied artificial intelligence Vol. 35; no. 15; pp. 2013 - 2036
• Main Authors: Khalilpour Darzi, Mohammad Rasoul, Khedmati, Majid, Niaki, Seyed Taghi Akhavan
• Format: Journal Article
• Language: English
• Published: Philadelphia Taylor & Francis 15.12.2021; Taylor & Francis Ltd; Taylor & Francis Group
• Subjects: Algorithms; Bayesian analysis; Classifiers; Customers; Data collection; Direct marketing; Marketing; Response rates
• ISSN: 0883-9514; 1087-6545; 1087-6545
• DOI: 10.1080/08839514.2021.1997226

Direct marketing identifies customers who buy, more probable, a specific product to reduce the cost and increase the response rate of a marketing campaign. The advancement of technology in the current era makes the data collection process easy. Hence, a large number of customer data can be stored in companies where they can be employed to solve the direct marketing problem. In this paper, a novel Bayesian method titled correlation-augment naïve Bayes (CAN) is proposed to improve the conventional naïve Bayes (NB) classifier. The performance of the proposed method in terms of the response rate is evaluated and compared to several well-known Bayesian networks and other well-known classifiers based on seven real-world datasets from different areas with different characteristics. The experimental results show that the proposed CAN method has a much better performance compared to the other investigated methods for direct marketing in almost all cases.