Experience Management with Task-Configurations and Task-Patterns for Descriptive Data Mining

Determining and instantiating an appropriate data mining task and method is often rather difficult, especially for inexperienced users. In this paper, we present a methodological approach for capturing, reusing, and generalizing experiences (task-configurations) for an easier selection and instantiation of appropriate methods and tasks in descriptive data mining. We show how the cases describing the task-configurations are structured, how they are retrieved, applied, and maintained. Furthermore, we present a process model that integrates the pro-