Advanced Data Mining and Analysis
In this release, Microsoft® SQL Server™ 2000 Analysis Services introduces a new feature, data mining, that integrates significant data analysis and prediction capabilities into Analysis Services. PivotTable® Service enables clients to interact with these new data mining features. For more information about data mining in Analysis Services, see Data Mining Models and Data Mining Columns.
PivotTable Service supports data mining by providing support services that are very similar to the services it provides for online analytical processing (OLAP). For example, PivotTable Service can create and maintain local data mining models just as it can create and maintain local cubes. To create a data mining model on an Analysis server, you must use Decision Support Objects (DSO). For more information about building mining models using DSO, see Data Mining Examples.
Two data mining algorithms are included with Analysis Services: Microsoft Decision Trees and Microsoft Clustering. The decision trees algorithm is based on the notion of classification. The clustering algorithm uses an expectation-maximization method to group records into clusters (or segments) that exhibit some similar, predictable characteristic. For more information, see Microsoft Clustering.
The following table describes topics that contain information about data mining in PivotTable Service. For detailed information about creating and using data mining models, including special functions for mining models, mining model XML format, and examples, see the OLE DB for Data Mining specification, available on the Microsoft OLE DB Web page at the Microsoft Web site.
| Topic | Description |
|---|---|
| Building a Local Data Mining Model | Describes the process of building local data mining models |
| Training a Local Data Mining Model | Describes how to process a local data mining model with training data |
| Predictions and Results of Data Mining | Describes how to run prediction queries against a data mining model and how to browse its contents |