Group Method of Data Handling

Gregory Ivakhnenko, National Institute for Strategic Studies Kyiv, Ukraina

The Group Method of Data Handling (GMDH) is self-organizing approach based on sorting-out of gradually complicated models and evaluation of them by external criterion on separate part of data sample.

Inductive GMDH algorithms gives possibility to find automatically interrelations in data, select optimal structure of model or network and increase the accuracy of existing algorithms. As input variables can be used any parameters, which can influence on the process. Linear or non-linear, probabilistic models or clusterizations are selected by minimal value of an external criterion. GMDH algorithms are rather simple and they get information directly from data sample.

This self-organizing approach is different from deductive methods or networks used commonly for modeling on principle. It has inductive nature - problems solution is based on sorting procedure by external criterion. The effective input variables, number of layers and neurons in hidden layers, optimal model structure are determined automatically. This is based on that fact that external criterion characteristic have minimum during complication of model structure. It was proved, that for inaccurate, noisy or small data can be found best optimal simplified model, accuracy of which is higher and structure is simpler than structure of usual full physical model. For real problems with noised or short data samples, simplified forecast models becomes more effective.

Group Method of Data Handling was applied in many countries for data mining and knowledge discovery, forecasting and  systems modelling, optimization and pattern recognition. Since 1968 many books, more than 230 doctoral dissertations were devoted to investigations in very different fields. Until now this approach was implemented in several commercial software products in USA and Germany.

The GMDH theory and source code of some algorithms was also published in 

"Self-Organising Data Mining" Mueller, J.-A., Lemke, F. 2000, ISBN 3-89811-861-4, Libri, Hamburg, http://www.knowledgeminer.net

"Inductive Learning Algorithms for Complex System Modeling",    Madala H.R. and Ivakhnenko A.G., 1994, ISBN: 0-8493-4438-7, CRC Press

"Self-organizing Methods in Modelling (Statistics: Textbooks and Monographs,vol.54)", Farlow, S.J. (ed.), 1984, ISBN: 0-8247-7161-3,      Marcel Dekker Inc.

GMDH books, articles and software can be found at http://www.niss.gov.ua/Center/articles/ .