Page 172 - DCAP606_BUSINESS_INTELLIGENCE

Unit 12: Understanding Data Mining Tools

The Data Mining Wizard in SQL Server Data Tools (SSDT) makes it very simple to create Notes
mining structures and mining models, using either relational data or multidimensional
data in cubes.

Notes In the wizard, you select data to use, and then request exact data mining techniques,
such as clustering, neural networks, or time sequence modelling.

Model viewers are supplied in both SQL Server administrations Studio and SQL Server
Data Tools (SSDT), for exploring your mining models after they are created.

Did u know? You can browse models using viewers tailored to each algorithm, or go
deeper into analysis by using the model content viewer.

The Prediction Query Builder is provided in both SQL Server administration Studio and
SQL Server Data Tools (SSDT) to help you create prediction queries.

Notes You can also test the accuracy of models against a holdout data set or external data,
or use cross-validation to assess the value of your data set.
SQL Server Management Studio is the interface where you organize living data mining
solutions that have been established to an example of Analysis Services. You can reprocess
structure and models to update the data in them.
SQL Server Integration Services includes tools that you can use to clean data, to automate
jobs such as creating predictions and updating models, and to create text mining solutions.

Let us discuss more about the data mining tools in SQL Server.

12.1.1 Data Mining Wizard

The Data Mining Wizard in Microsoft SQL Server Analysis Services starts every time when you
add a new mining structure to a data mining project. It helps you choose a data source and set up
a data source model view that defines the data to be used for analysis, and then helps you create
an initial model.

Starting the Data Mining Wizard

To use the Data Mining Wizard, you must have opened a solution in SQL Server Data Tools
(SSDT) that contains at least one data mining or OLAP project. Follow these steps:
If your solution is ready for data mining, you can right-click the Mining Structures node in
Solution Explorer and select New Mining Structure to start the wizard.

If your solution does not contain any existing projects, you can add a new data mining
project. To add a new project, from the File menu, select New, and then select Project. Then
chose the template, Analysis Services Multidimensional and Data Mining Project.
Relational or OLAP Mining Model

The next decision to make is whether to use a relational data source, or to use an OLAP mining
model.

LOVELY PROFESSIONAL UNIVERSITY 167