Počet záznamů: 1
Performance of classification confidence measures in dynamic classifier systems
- 1.
SYSNO ASEP 0423771 Druh ASEP J - Článek v odborném periodiku Zařazení RIV J - Článek v odborném periodiku Poddruh J Článek ve WOS Název Performance of classification confidence measures in dynamic classifier systems Tvůrce(i) Štefka, D. (CZ)
Holeňa, Martin (UIVT-O) SAI, RIDZdroj.dok. Neural Network World. - : Ústav informatiky AV ČR, v. v. i. - ISSN 1210-0552
Roč. 23, č. 4 (2013), s. 299-319Poč.str. 21 s. Jazyk dok. eng - angličtina Země vyd. CZ - Česká republika Klíč. slova classifier combining ; dynamic classifier systems ; classification confidence Vědní obor RIV IN - Informatika CEP GA13-17187S GA ČR - Grantová agentura ČR Institucionální podpora UIVT-O - RVO:67985807 UT WOS 000325193300003 EID SCOPUS 84885585402 DOI 10.14311/NNW.2013.23.019 Anotace Classifier combining is a popular technique for improving classification quality. Common methods for classifier combining can be further improved by using dynamic classification confidence measures which adapt to the currently classified pattern. However, in the case of dynamic classifier systems, the classification confidence measures need to be studied in a broader context as we show in this paper, the degree of consensus of the whole classifier team plays a key role in the process. We discuss the properties which should hold for a good confidence measure, and we define two methods for predicting the feasibility of a given classification confidence measure to a given classifier team and given data. Experimental results on 6 artificial and 20 real-world benchmark datasets show that for both methods, there is a statistically significant correlation between the feasibility of the measure, and the actual improvement in classification accuracy of the whole classifier system; therefore, both feasibility measures can be used in practical applications to choose an optimal classification confidence measure. Pracoviště Ústav informatiky Kontakt Tereza Šírová, sirova@cs.cas.cz, Tel.: 266 053 800 Rok sběru 2014
Počet záznamů: 1