Dataprix

Datamining, is the no trivial process of discovering valid patterns, new, potentially useful and comprehensible inside a group of data, as the definition of Piatetsky-Shapiro published in the magazine "AI Magazine".
To simplify, we could say that data mining treats to extract knowledge from data.
By means of a series of processes applied in different phases on the data in gross, and defined by an expert that know the meaning of these data, and have clear the aims that pursues, can extract relations between these data, discover unseen patterns and build models that describe this knowledge.
The phases by which would have to happen this process of finding of knowledge are the following:

- Definition of the datamining task.
¿What objective pursue ?
- Data selection
- Data preparation
- Application of datamining processes on the treated data
- Evaluation and interpretation of the model obtained
- Integration of the results in the information systems

It is a continuous process, and can feature of different iterations, where the results of an iteration feeds the start of the following.

In these slides the case of study of a bank appears that considers the necessity of a greater knowledge of its clients to be able to define its strategies of business suitably. Thanks to the use of tools of Business Intelligence, in particular of Data Warehouse and Data Mining, and to the definition of clear objectives of business, this bank could analyze the behavior of its clients, segment them, make decisions estraté gicas based on this behavior, to make predictions and to analyze the results of the application of these decisions, being valued therefore the return of the investment. The presentation is structured in the following sections: - The economy of the business and the management of clients - Modelamiento of value - Modelamiento of potential - Segmentation - Modelamiento of desertion - précticas Applications - Results