Entropy and Quantity of Information in Technical Designations
V.G. Kryzhanovskyi · Visnyk of Vinnytsia Politechnical Institute · 2023
Conventional designations of integrated microcircuits are considered as an example of classification and abbreviated name (code) of technical products to answer the question: Why do they say, that some designation systems are "more informative?". Do such notations contain more information compared to other systems? Such tasks are closely related to the tasks of machine learning and the construction of the "semantic web". Based on the algebraic approach and set theory, the characteristics of the entropy of the classification of designations are considered and it is shown that the entropy of such a coded designation is less than that of an arbitrary system of recording technical characteristics, which is explained by the positional structure of the designation and, accordingly, the lower power of the sets that make up a specific designation. Based on the approach of informational algebra, it is confirmed that the establishment in the technical notation of the atomic structure of the sets to which the technical characteristics correspond, really corresponds to the mathematical definition of a more informative structure. Based on the mathematical theory of hints, the structure of the technical designation is analyzed and the possibility of obtaining additional information, for example, relationships between different groups of technical parameters, is indicated. It will be obtained as a result of questions clarifying the interpretation of existing answers. This is a consequence of the property of hint entropy, which has two components — the Shannon entropy and the generalized Hartley measure, which correspond to probabilistic information about the true interpretation of the answer in the set and relational information about the true answer about some type of integrated circuit parameters. Technical notation turns out to be an effective example on which the considered mathematical theories can be applied and accordingly can be an example of a code that, on the one hand, can be understood by a person, and on the other hand, can be used in machine information processing systems.