HEURISTIC METHOD FOR FORMING CLASSES USING ALGORITHMS FOR PATTERN RECOGNITION

Filip Iliev

Scientific paper ID 1028 : 2014/3

HEURISTIC METHOD FOR FORMING CLASSES USING ALGORITHMS FOR PATTERN RECOGNITION

Filip Iliev

The present paper proposes a heuristic method for forming classes necessary for recognition of objects and events, in which is used desirability function. Its application does not depend on the probability distribution, and through him are characterized extreme values of the aggregation, multiplied upon the statistical sample which is very comfortably for practical purposes, when it is necessary to distinguish between two classes containing a protected area. The method combines the importance of physical and psychological parameters through a certain scale of preference and is particularly suitable in the presence of expert assessments.

• Reload article • Open/Download as PDF

разпознаване на образи формиране на класове функция на желателността (desirable).pattern recognition formation of classes desirability function.Filip Iliev

BIBLIOGRAPHY

[1] Harrington Е.С., Industrial Quality Control, 1965, 21, N10.

[2] Gumbel E.J., Statistics of Extremes. Columbia University Press, New York, 1958.

[3] Nikolov H., S. Ruseva, Application of Nonparametric Bayesian Classifier to Remote Sensing Data. Information Technologies and Control, No.2, 2004, p.8-12.

[4] Nikolov H., S. Ruseva, Feature selection methods in remotely sensed multispectral data classification. 4-th National Geophysical Conference, Sofia, 4-5 October 2004.

[5] Smirnov H.B., I.V. Dunin-Barkovskiy, Kurs teorii veroyatnostey i matematicheskoy statistiki dlya tehnicheskih prilozheniy. M., 1965.
( [5] Смирнов H.B., И.В. Дунин-Барковский, Курс теории вероятностей и математической статистики для технических приложений. М., 1965. )

[6] Tsypkin Ya.Z., Informatsionnaya teoriya identifikatsii. M., Nauka, 1995.
( [6] Цыпкин Я.З., Информационная теория идентификации. М., Наука, 1995. )