Introduction
 - general data analysis methodology
   * object oriented technology
   * data analysis stages : preprocessing, classification
   * search for a robust combination of methods
Algorytmy
  * wprowadzenie - minimalizacja funkcji kosztw
  * model selection criteria
    - balance - a special case of cost matrix
 - klasyfikacja
   * kNN
     - algorithm
     - analysis in the context of the contest
       (distances atrongly depend on the feature set, standardization etc.)
   * SVM
   * SSV Tree
 - selekcja z kontekstem jak pomoc klasyfikatorom
     filters and wrappers
   * CC based
   * SSV based
   * Wrapper
   * Selector Committee	???
   * Sequential Feature Selection ???
   * Information theory based filters
     - not used finally, similar results on the whole datasets
 - other useful data transformations
   * Std
   * PCA
 - model validation methods
   * supervised and unsupervised data preparation
   * CV Committees
   * Meta Search
    - important to have a flexible system
Challenge Results
 - Introduction
   * each dataset of different kind - different methods
 - Arcene
 - Dexter
 - Dorothea
 - Gisette
 - Madelon 
Summary
 - amazing, how fruitful can CC be
 - data analysis dangers
   * overfitting by models easily validated, overfitting 
   by preprocessing more difficult
 