How to locate the target user is a very important topic in any business unit, especially in the case of limited budget, how to get the maximum benefit of the activity, the target user's positioning is an important means.
This article describes how to locate target users based on historical information through the data mining capabilities in SQL Server Analysis Services (SSAS).
Microsoft SQL Server provides a complete solution for the data platform, he is not just a data engine, but also provides a lot of data tools and services, with the help of Analysis Services, combined with the massive historical data information in the business system, SQL Server can help us discover the patterns and rules, Therefore, the target data can be forecasted and analyzed.
In practice, different mining models for different problem scenarios, the same problem model may have multiple models are suitable to solve the problem, so that the problem to solve a lot of contrast, so that we can predict the accuracy of each model to choose an optimal model.
In this paper, three kinds of mining models are used to compare the target user location, such as decision Tree, Bayesian and clustering algorithm.
The SQL Server version used in this article is 2012, and the sample database is the Data Warehouse for Adventure Works, and you can refer to my essay on how to obtain and deploy this sample Data warehouse:
Http://www.cnblogs.com/aspnetx/archive/2013/01/30/2883831.html
First, build the data mining project and open the SQL data TOOLS, which is the shell of Visual Studio 2010.
Create a new analysis service for multiple and data mining projects.
When the project is established, create a new data source connection.
Select the deployed Adventure Works DW connection.