Geodesic
Search for the optimal metric in classification problems with ordinal features
Žurnal vyčislitelʹnoj matematiki i matematičeskoj fiziki, Tome 50 (2010) no. 3, pp. 585-592 
G. V. Iofina. Search for the optimal metric in classification problems with ordinal features. Žurnal vyčislitelʹnoj matematiki i matematičeskoj fiziki, Tome 50 (2010) no. 3, pp. 585-592. http://geodesic.mathdoc.fr/item/ZVMMF_2010_50_3_a16/
@article{ZVMMF_2010_50_3_a16,
     author = {G. V. Iofina},
     title = {Search for the optimal metric in classification problems with ordinal features},
     journal = {\v{Z}urnal vy\v{c}islitelʹnoj matematiki i matemati\v{c}eskoj fiziki},
     pages = {585--592},
     year = {2010},
     volume = {50},
     number = {3},
     language = {ru},
     url = {http://geodesic.mathdoc.fr/item/ZVMMF_2010_50_3_a16/}
}
TY  - JOUR
AU  - G. V. Iofina
TI  - Search for the optimal metric in classification problems with ordinal features
JO  - Žurnal vyčislitelʹnoj matematiki i matematičeskoj fiziki
PY  - 2010
SP  - 585
EP  - 592
VL  - 50
IS  - 3
UR  - http://geodesic.mathdoc.fr/item/ZVMMF_2010_50_3_a16/
LA  - ru
ID  - ZVMMF_2010_50_3_a16
ER  - 
%0 Journal Article
%A G. V. Iofina
%T Search for the optimal metric in classification problems with ordinal features
%J Žurnal vyčislitelʹnoj matematiki i matematičeskoj fiziki
%D 2010
%P 585-592
%V 50
%N 3
%U http://geodesic.mathdoc.fr/item/ZVMMF_2010_50_3_a16/
%G ru
%F ZVMMF_2010_50_3_a16

Voir la notice de l'article provenant de la source Math-Net.Ru

A method for finding the optimal distance function for the classification problem with two classes in which the objects are specified by vectors of their ordinal features is proposed. An optimal distance function is sought by the minimization of the weighted difference of the average intraclass and interclass distances. It is assumed that a specific distance function is given for each feature, which is defined on the Cartesian product of the set of integer numbers in the range from $0$ to $N-1$ and takes values from $0$ to $M$. Distance functions satisfy modified metric properties. The number of admissible distance functions is calculated, which enables one to significantly reduce the complexity of the problem. To verify the appropriateness of metric optimization and to perform experiments, the nearest neighbor algorithm is used.

[1] Zhuravlev Yu. I., Ryazanov V. V., Senko O. V., “Raspoznavanie”. Matematicheskie metody. Programmnaya sistema. Prakticheskie primeneniya, Fazis, M., 2006

[2] Iofina G. V., “Vybor nailuchshei metriki v algoritme raspoznavaniya po blizhaishemu sosedu”, Sb. tr. 49-i nauchn. konf. (MFTI), MFTI, M., 2006, 266–267

[3] Galeev E. M., Tikhomirov V. M., Kratkii kurs teorii ekstremalnykh zadach, MGU, M., 1989

[4] Sadovnichii V., Grigoryan A., Konyagin S., Zadachi studencheskikh matematicheskikh olimpiad MGU, MGU, M., 1987, 87; 88

[5] Newman D. J., Hettich S., Blake C. L., Merz C. J., UCI repository of machine learning databases, Techn. Rept. Inst. California, Irvin Dept. Inform. and Comp. Sci., 1998

[6] Dietterich T. G., “Approximate statistical tests for comparing supervised classification learning algorithms”, Neural Comput., 1998, no. 10, 1895–1924 | DOI