Data Mining Tutorials (Analysis Services)
Updated: February 3, 2016
Applies To: SQL Server 2016 Preview
Microsoft SQL Server Analysis Services makes it easy to create data mining solutions using wizards and integrated visualizations. Particularly if you are new to machine learning, the tools in Analysis Services are an easy way to design, train, and explore data mining models. The data in your models can be stored in a cube, relational database, or any other source support by Analysis Services. After creating a model, you can put it into production by accessing the model to create predictions using prediction multiple clients, including Integration Services and ASP.NET.
SQL Server 2016 is in a pre-release version and the data mining tutorials have not been updated. We recommend that you reference the versions of the tutorials created for SQL Server 2014. The steps should be identical.
In This Section
This tutorial walks you through a targeted mailing scenario. It demonstrates how to use the data mining algorithms, mining model viewers, and data mining tools that are included in Analysis Services. You will build three data mining models to answer practical business questions while learning data mining concepts and tools.
This tutorial contains a collection of lessons that introduce more advanced data mining concepts and techniques. The scenarios include these model types:
market basket analysis
neural networks and logistic regression
The lessons are independent and can be done in any order, but you should have a basic knowledge of how to build data sources.
Advanced concepts covered in these lessons include the use of nested tables, cross-prediction, custom data source views and named queries, and filtering in data mining queries. You will also gain proficiency in using the prediction query tools that are included in Analysis Services.
The Data Mining Extensions (DMX) query language has syntax like that of SQL but can be used to create, query, and manage predictive models stored in Analysis Services.
These tutorials demonstrate how to create a new mining structure and mining models by using the DMX language, and how to create DMX prediction queries for use in applications.