Realtime Detection from Customer’s Behavior Sequence – Explore a Smart Customer Maintenance Algorithm

• Activity Number: 500 - Statistical Challenges and Recent Advances in Finance and Business Analytics
• Type: Topic Contributed
• Date/Time: Wednesday, July 31, 2019 : 10:30 AM to 12:20 PM
• Sponsor: Business and Economic Statistics Section
• Author(s): Mingfei Li*
• Companies: Bentley University
• Keywords: Business Analytics; Statistical learning; Sequencial prediction; Real time computing; Marketing analytics

Abstract

Motivated by practical marketing challenges on big data, we are interested in developing a new statistical learning algorithm from customer’s behavior sequentially, to help real-time customer maintenance. Our past study explored the long-term forecasting and proposed the idea of adaptive windows for online learning for detection. This study we will put a focus on the detection of customer’s behavior pattern change. We will explore this topic with both simulations and real business data and discuss the balance between forecast and detection function for the new algorithm.