OLAP-Based Data Mining for Business Intelligence Applications in Telecommunications and E-commerce

• Authors: Qiming Chen;Umeshwar Dayal;Meichun Hsu
• Venue: DNIS '00 Proceedings of the International Workshop on Databases in Networked Information Systems
• Year: 2000

Business intelligence applications require the analysis and mining of large volumes of transaction data to support business managers in making informed decisions. In a telecommunication network, hundreds of millions of call detail records are generated daily. Business intelligence applications such as fraud detection and churn analysis require the collection and mining of these records on a continuous basis. Similarly, electronic commerce applications require the analysis of millions of shopping transaction records daily to guide personalized marketing, promotional campaigns, and fraud detection. An important component of many of these applications is customer profiling, which aims to extract patterns of behavior from a collection of transaction records, and the comparison of such patterns. The high data volumes and data flow rates pose serious scalability and performance challenges. We show how a scalable data-warehouse/OLAP framework for customer profiling and pattern comparison can meet these performance requirements. Also, since it is important in many business intelligence applications to collect and analyze transaction records continuously, rather than in batches, we show how to automate the whole operation chain, including data capture, filtering, loading, and incremental summarization and analysis.