H.A. Kravtsov, V.I. Koshel
Èlektron. model. 2017, 39(5):59-70
https://doi.org/10.15407/emodel.39.05.059
ABSTRACT
The methods of artificial intelligence used to cataloguize information require the existence of classifications or taxonomies for determining class affiliation of subjects, phenomena, actions etc. However, for correct solving of classification problems it is necessary that all the used classifications/ taxonomies were correct. The authors consider the notion «classification correctness” and investigate the possibility to identify some errors of dividing by using the theory of classifications calculus. The authors also propose visual and heuristic approaches for detecting the following errors: dividing with remainder terms, controversial dividing and dividing jump.
KEYWORDS
classification, correctness, division errors, measure of difference, fixed class, normalized measure.
REFERENCES
1. Kravtsov, H.A. (2016), “Measure of difference between classifications”, Elektronnoe modelirovanie, Vol. 38, no. 4, pp. 81-97.
2. Kravtsov, H.A. (2016), “Model of computations over classifications”, Elektronnoe modelirovanie, Vol. 38, no. 1, pp. 73-87.
3. Berztiss, A.T. (1974), Struktura dannykh [Data structure], Statistika, Moscow, USSR.
4. Adamek, J., Herrlich, H. and Strecker, G.E. “Abstract and concrete categories. The joy of cats”, available at: http://katmat.math.uni-bremen.de/acc/acc.pdf. (accessed June, 2017).
5. Korotkov, E.M. (2004), Issledovanie system upravleniya [Study of control systems], DeKA, Moscow, Russia.
6. Ivlev, Yu.V. (2008), Logika [Logic], TK Velbi, Prospekt, Moscow, Russia.
7. Bukvy!: Pravila deleniya v logike i oshibki deleniya [Letters! Dividing rules in the logic and dividing errors], available at: http://bukvi.ru/pravo/logika/pravila-deleniya-v-logike-i-oshibkiv-delenii.html. (accessed June, 2017).