Data mining exploits the knowledge that is held in the enterprise data store by examining the data to reveal patterns that suggest better ways to produce profit, savings, higher-quality products, and greater customer satisfaction. Just as the lines on our faces reveal a history of laughter and frowns, the patterns embedded in data reveal a history of, for example, profits and losses. The retrieval of these patterns from data and the implementation of the lessons learned from the patterns are what data mining and knowledge discovery are all about.
This book will appeal to people who have come to depend upon Microsoft to provide a high-performance and economical point of entry for an ever-increasing range of computer applications and who sense the potential value of pursuing data mining approaches to support business intelligence initiatives in their enterprises. Traditional producers and consumers of business intelligence products and processes, especially OLAP (On-Line Analytical Processing), will also be attracted by this information. Most business intelligence vendors, especially Microsoft, recognize that business intelligence and data mining are different facets of the same process of turning data into knowledge. SQL Server 7, released late in 1998, introduced SQL Server 7 OLAP services, thus providing a built-in OLAP reporting facility for the database. In the same manner, SQL Server 2000 provides built-in data mining services as a fundamental part of the database. Now, both these important forms of business reporting will be available as core components of the database functionality; further, by providing both sets of facilities in a common interface and platform, Microsoft has taken the first step in providing a seamless integration of the various methods and metaphors of business reporting so that one simple, unified interface to the knowledge contained in data is provided. Whether that knowledge was delivered on the basis of an OLAP technique or data mining technique is irrelevant to most users, and now it will be irrelevant in a unified SQL 2000 framework.