Creating data mining models (modeling)

Intelligent Miner® provides modeling technology as Db2 extenders. This modeling technology is called via an SQL API.

By using the SQL API of Intelligent Miner, you can use the following mining functions to develop analytic PMML models that are stored in Db2 tables:

  • Association rules
  • Sequence rules
  • Clustering
  • Regression
  • Classification
  • Time Series Forecasting

You can process these PMML models for scoring or visualization.

Ease of use

The following features support ease of use for model creation:

Easy mining procedures
The Easy Mining procedures provide high-level task-oriented SQL procedures. With the Easy Mining procedures, you can perform data mining efficiently and successfully in a business context without the need of in-depth data mining skills. You only need to be familiar with the appropriate business area where you want to apply data mining. There are Easy Mining procedures for common mining tasks and for basic mining steps. The Easy Mining procedures are described in IBM® warehousing in Db2®: Data Mining with Easy Mining procedures.
Self-configuration and input formats
Intelligent Miner improves self-configuration and self-optimization in the mining algorithms:
  • You need not define tuning parameters for the mining functions. The mining functions automatically define suitable sets of tuning parameters that match the characteristics of the training data.
  • The Association rules mining function and the Sequence Rules mining function can read the groups of items out of tables with different table layouts.
  • The Association rules mining function and the Sequence Rules mining function can automatically adjust the rule-filter criteria to create models with a user-defined number of rules.
  • The Association rules mining function and the Sequence Rules mining function support different item formats including a multi-value item format. With the multi-value format, you can place an arbitrary number of single items in a column value of an item column.
Automatic variable selection
Intelligent Miner provides automatic variable selection to control field importance for the mining process. The algorithms detect which fields to use for mining and perform some data conditioning (such as discretization or pivoting in Associations) or internal transformations for date, time, and timestamp without manual interaction.