statistical pattern recognition,
intellectual data mining.
Research of statistical stability of pattern classifiers in the condition of small sample size. Studying the probabilistic properties of decision functions for discrete recognition problem; finding the dependences between sample size, number of values of discrete characteristics, empirical error and the expected probability of error. Developing of algorithms for decision trees construction in problems of pattern recognition, clustering, regression analysis, time series analysis. Application of the developed data mining methods in applied researches.
Graduated from Faculty of Mathematics and Mechanics of Novosibirsk State University in 1986 (department of theoretical cybernetics). Ph. D. thesis was defended in 1996, Doctor Nauk thesis was defended in 2007. A list of journal articles contains 49 titles.
Award from International Academic Publishing Company "Nauka/Interperiodica" for series of papers on results in theory and practice of constructing decision recognition functions based on the analysis of empirical information (1998).
Lbov G.S., Berikov V.B., Stability of decision functions in problems of pattern recognition and analysis of heterogeneous information, Sobolev institute of mathematics publishing, 2005 (in Russian)
Berikov V. B., “An approach to the evaluation of the performance of a discrete classifier”, Pattern Recogn. Letters, 23:1-3 (2002), 227–233
Berikov V.B., Lbov G.S., “Bayes estimates for recognition quality on finite sets of events”, Doklady Mathematics, 71:3 (2005), 327–330
Berikov V.B., “Grouping of Objects in a Space of Heterogeneous Variables with the Use of Taxonomic Decision Trees”, Pattern Recognition and Image Analysis, 21:4 (2011), 591–598
Berikov V.B., “A latent variable pairwise classification model of a clustering ensemble
\ibyy C. Sansone, J. Kittler, and F. Roli (Eds.)”, Multiple Classifier Systems, 2011, Lecture Notes on Computer Science, 6713, Springer, Berlin, 2011, 279–288