The paper proposes a solution to the problem classification by calculating the\nsequence of matrices of feature indices that approximate invariants of the data\nmatrix. Here the feature index is the index of interval for feature values, and\nthe number of intervals is a parameter. Objects with the equal indices form\ngranules, including information granules, which correspond to the objects of\nthe training sample of a certain class. From the ratios of the information granules\nlengths, we obtain the frequency intervals of any feature that are the\nsame for the appropriate objects of the control sample. Then, for an arbitrary\nobject, we find object probability estimation in each class and then the class of\nobject that corresponds to the maximum probability. For a sequence of the\nparameter values, we find a converging sequence of error rates. An additional\neffect is created by the parameters aimed at increasing the data variety and\ncompressing rare data. The high accuracy and stability of the results obtained\nusing this method have been confirmed for nine data set from the UCI repository.\nThe proposed method has obvious advantages over existing ones due\nto the algorithm�s simplicity and universality, as well as the accuracy of the\nsolutions.
Loading....