Steps to complete the high level of Data Mining

Developing and understanding domain Creating the target data set by selecting a proper set of variables or data sample Data cleaning and preprocessing Data reduction and projection Selecting the data mining algorithm/s Data mining Interpreting mined pattern Consolidating discovered knowledge

One of the most difficult tasks is to choose the right data mining technique

Two main parameters that the final choice depends basically on:
The main goal of problem to be solved The structure of available data

Knowledge models

No response variable

With response variable

Descriptive models

Associative models

Discriminant models

Predictive models

CBR (CASE BASED REASONIG)

Problem solving methodology that tries to solve new problems bye re-using specific past experiences stored in a cased based and its proposition is that new problem can be solved by reusing the solutions to similar problems that have been solved in the past.

Approach allows the contruction of dynamic system able to refine and adapt the suggestions over time

CBR APPROACH
Case description: describing the type of problem the data miner wants to solve Case solution: a set of DM techniques with its own configuration parameter

FOUR MAIN PHASES ABOUT CBR

Retrieve Reuse Revise Retain