Data Mining Programming
If you find that the built-in tools and viewers in Analysis Services do not meet your requirements, you can extend the power of Analysis Services by coding your own extensions. In this approach, you have two options:
Microsoft SQL Server 2014 Analysis Services (SSAS) supports XML for Analysis (XMLA) as a protocol for communication with client applications. Additional commands are supported by Analysis Services that extend the XML for Analysis specification.
Because Analysis Services uses XMLA for data definition, data manipulation, and data control support, you can create mining structures and mining models by using the visual tools provided by SQL Server Data Tools (SSDT), and then extend the data mining objects that you have created by using Data Mining Extensions (DMX) and Analysis Services Scripting Language (ASSL) scripts.
You can create and modify data mining objects entirely in XMLA scripts, and run prediction queries against the models programmatically from your own applications.
Analysis Management Objects (AMO)
Analysis Services also provides a complete framework that enables third-party data mining providers to integrate the data mining objects into Analysis Services.
You can create mining structures and mining models by using AMO. See the following samples in CodePlex:
Connects to the SSAS instance you specify and lists all server objects and their properties, including mining structure and mining models.
AMO Simple Sample
The AS Simple Sample covers programmatic access to most major objects, and demonstrates metadata browsing, and access to the values in objects.
The sample also demonstrates how to create and process a data mining structure and model, as well as browse an existing data mining model.
You can use DMX to encapsulate command statements, prediction queries, and metadata queries and return results in a tabular format, assuming you have created a connection to an Analysis Services server.