Languages of publication
Three methods for distinctive (independent) features determination were designed and implemented in computer programs. Two of them were based on the assumption that the maximum number of the dependent variable’s values cannot exceed the number of the points of the variable space. Using the third method, some functional relations between a priori chosen dependent features and the sets of a priori chosen independent features were studied. The main purpose of another method was to determine the minimum sets of features necessary to distinguish individual objects in a given set. It allowed to evaluate features’ “load” (how often they were used in the task of object’s distinguishing). In this way, any complex functional relations of any number of categorical features which can take any number of values can be determined. Possible applications vary from optimizing phones’ description in the articulatory space, to the optimal selection of graphical features of different alphabets’ signs, and database normalization.
Publication order reference