Confidence of Classification and its Application to Classifier Aggregation

Kód výsledku v IS VaVaI
<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F68407700%3A21340%2F07%3A04137650" target="_blank" >RIV/68407700:21340/07:04137650 - isvavai.cz</a>
Výsledek na webu
—
DOI - Digital Object Identifier
—

Jazyk výsledku
angličtina
Název v původním jazyce
Confidence of Classification and its Application to Classifier Aggregation
Popis výsledku v původním jazyce
Classifier combining is a succesful method for improving the quality of classification. In this paper, we introduce the concept of confidence of classification and define two confidence measures -- the local accuracy and the local diversity. We propose algorithms for classifier aggregation which utilize the concept of confidence. We then compare the performance of these algorithms with several state-of-the-art classifier combining techniques. Results of these benchmark tests show that by incorporating confidence into classifier aggregation algorithms, the state-of-the-art methods can be improved.
Název v anglickém jazyce
Confidence of Classification and its Application to Classifier Aggregation
Popis výsledku anglicky
Classifier combining is a succesful method for improving the quality of classification. In this paper, we introduce the concept of confidence of classification and define two confidence measures -- the local accuracy and the local diversity. We propose algorithms for classifier aggregation which utilize the concept of confidence. We then compare the performance of these algorithms with several state-of-the-art classifier combining techniques. Results of these benchmark tests show that by incorporating confidence into classifier aggregation algorithms, the state-of-the-art methods can be improved.

Rok uplatnění
2007
Kód důvěrnosti údajů
S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Podobné výsledky(10)