Soft cyclic data encoding using a quasi-fuzzy measure

In the article, the authors propose a new numerical method for representing quantitative variables that have several qualitative, cyclically repeating attributes. Such variables will be represented by a vector whose elements reflect the degree to which the value of the variable corresponds to each of the qualitative features. To encode quality features, the authors propose using special functions that have the specified properties. The developed procedure is called quasi-fuzzy coding. As a result, the coding vector is able to adequately represent the original variable, reflecting its qualitative properties. The article presents the results of computational experiments that confirm the effectiveness of the proposed method.