Researcher's view on the MiningMart project

Within the data mining process considerable time is spent for pre-processing the data. Practical experiences have shown that the time spent on preprocessing can take from 50% up to 80% of the entire data mining process when using the traditional attribute-value learners. Thats why preprocessing is the key issue in data analysis. The time is spend for:

• Choosing the learning task
• Sampling
• Feature generation, extraction, and selection
• Data cleaning
• Model selection or tuning the hypothesis space
• Defining appropriate evaluation criteria

Experienced users can apply any learning system successfully to any application, since they prepare the data well. The representation of examples and the choice of a sample determines the applicability of learning methods. A chain of data transformations (learning steps or manual preprocessing) delivers the desired result. Experienced users remember prototypical successful transformation/learning chains.

The MiningMart Approach

The basic idea is to store best practice cases of preprocessing chains that where developed by experienced users. The data is described on the meta level and is presented in application terms. MiningMart users choose a case and apply the corresponding transformation and learning chain to their application.

The project has developed new techniques that support user-guided representation adjustment as well as techniques that automatically select or change representations:

Next...

Old version of the MiningMart website