The Criterion of Congruence in the Theory of Selforganization. N.A. Ivakhnenko. IWIM, Prague, 2007.
Article (in pdf)
Abstract. The criterion of congruence is one of the selection ones. It selects the best solutions both for clusterization and modeling of processes.. It consists of the comparison of both clusterisions on two square arrays of specially organized points. Such arrays are named “faces” of given clusterisation ( in Russian – “Litso”). These faces help to compare arrays with various quantity of clusters and various quantity points in them simultaneously, and because this criterion is called as a congruent one. The congruent criterions are called such ones, that take in the consideration two or more special examine qualities of objects simultaneously. For our tasks such objects serve clusters and their contents. So in the case of the algorithms of Self-organizations this is used firstly for finding the multitudes of arguments for two arrays, selecting better ones for electing the criterion. Then knowing a few better multitudes of our arguments, using already knowing procedures in the first part, find the multitudes of arguments for the full our massive. Our criterion differs from others and its results of action differs also, because you can compare only your finish results for its using. For the modeling of processes you must hold on the same structures of the algorithm’s steps, but for founding of need multitudes you must not take the function value. And only the end step use the ordinary function analysis. The propose criterion will open area of other selections ones in the future.
Keywords. Inductive modelling, GMDH, criterion of congruence, clusterization, self-organization.
References.
- Ivakhnenko A.G., Stepashko V.S. Pomekhoustoychivost’ modelirovaniya (Noise - immunityof modeling) .Naukova Dumka, Kiev, 1985.
- Ivakhnenko A.G., Yurachkovsskiy. Modelirovaniye slzhnykh system po eksperimentalnym dannym ( “Modeling of Complex Systems for Experimental Data). Radio I svyaz, Moscow,1987
- Madala H.R., Ivaknenko A.G. Inuctive learning algorithms for complex system modeling, CRC., Press., 1994.
- Vorontsov K.V. About combinatorial exit to the learning algorithms. Reception for the eleventh United conference in Putscino, Russian, 2003.
