Discrete and Continuous Models and Applied Computational ScienceDiscrete and Continuous Models and Applied Computational Science2658-46702658-7149Peoples' Friendship University of Russia8368Research ArticleComparative Study of Cluster and Neural Network Methods in the Problem of Protein Structure AnalysisBaranovD ALaboratory of Information TechnologiesDmitriyBaranof@gmail.comOsoskovG ALaboratory of Information Technologiesososkov@jinr.ruBaranovA Aandbar91@yandex.ruJoint Institute for Nuclear ResearchMoscow State Institute of Radio Engineering, Electronics and Automation15022014223423808092016Copyright © 2014,2014This work continues the previous study where the important problem of automatization of differentiation methods of the genetic protein structures according to their electrophoretic spectrums (EPS) was considered. The multicriterion problem of the agriculture cultivar identification by their spectra caused the idea of its solution by an artificial neural network (ANN) trained on an expert data base. In the given paper peculiarities of the neural net use as well as the purposefulness of cluster analysis applications for the EPS classifying are studied. A special model of multidimensional vectors adequately imitating the most essential characteristics of real data obtained after EPS digitalization, denoising and normalization is developed. A numerical experiment is fulfilled on such simulated data stream to study the influence of contamination and distortion factors on the ANN efficiency in order to suppress those factors and improve ANN functioning. Various methods of cluster analysis are also applied to simulated multidimensional data as either an ANN alternative or more soundly as a prior stage of a coarse data classification in some set of detached cultivar groups to be classified next by ANN.artificial neural networksclassificationclusterizationgenetic analysiselectrophoretic spectraискусственные нейронные сетиклассификациякластеризациягенетический анализопределение сортовой принадлежностиэлектрофоретический спектр[Haykin S. Nueral Networks. A Comprehensive Foundation. - New Jersey: Prentice Hall, 2006.][Ososkov G. A., Baranov D. A. Feature Extraction for Data Input to Neuro-Classifiers // Mathematical Modeling and Computational Physics (MMCP 2009): Book of Abstract of the International Conference (Dubna, July 7-11). - 2009. - Pp. 110-111.][Ososkov G. A., Baranov D. A. Extraction of Data Features for Neuro-Classifier Input // Bulletin of PFUR. Series “Mathematics. Information Sciences. Physics”. - 2010. - No 3. - Pp. 142-148.][Ruanet V. V., Kudryavtsev A. M., Dadashev S. Y. The Use of Artificial Neural Networks for Automatic Analysis and Genetic Identification of Gliadin Electrophoretic Spectra in Durum Wheat // Russian Journal of Genetics. - 2001. - Vol. 37, No 10. - Pp. 1207-1209.][Duran B. S., Odell P. L. Cluster Analysis. - New York: Springer Verlag, 1947.]