Methods of true data mining model selection - with experimental results. Pavel Kordik, Oleksandra Bulgakova

Abstract. This work presents the modeling results of different real noisy data (nalada: humanities, spirals_1 and spirals_2: too complex data, motol_brain: motol hospinal neurosurgery, boshouse: house prices and also artificial data), using intellectual computing – combinatorial group method of data handling (combi GMDH) and Group of adaptive models evolution (GAME) method. All this data, you can find in [1]. The goal of our work is to get the best possible result on such noisy data and to compare the results of particular methods. Also, we will make to attempt to combine these two methods (Game_Combi_GMDH) for the receipt of more high-quality prediction.

Keywords. Data mining, prediction, real noisy data, group method of data handling (GMDH), group of adaptive models evolution (GAME), inductive modeling.

References.

1. Files with data sets (open source):  http://neuron.felk.cvut.cz/game/data/.

2. GMDH algorithms: describe, application, examples.  http://www.gmdh.net/GMDH_com.htmv