Comparing classifiers for pronunciation error detection
conference paper
Providing feedback on pronunciation errors in computer assisted language learning systems requires that pronunciation errors be detected automatically. In the present study we compare four types of classifiers that can be used for this purpose: two acoustic-phonetic classifiers (one of which employs linear-discriminant analysis (LDA)), a classifier based on cepstral coefficients in combination with LDA, and one based on confidence measures (the so-called Goodness Of Pronunciation scores). The best results were obtained for the two LDA classifiers which produced accuracy levels of about 85-93%. Index Terms: Computer Assisted Pronunciation Training (CAPT), pronunciation error detection, acoustic-phonetic classifiers, Goodness Of Pronunciation (GOP).
Topics
Acoustic-phonetic classifiersComputer Assisted Pronunciation Training (CAPT)Goodness of Pronunciation (GOP)Pronunciation error detectionAcousticsClassifiersDiscriminant analysisError detectionErrorsLearning systemsLinguisticsSpeechSpeech recognitionAccuracy levelsAcoustic-phonetic classifiersCepstral coefficientsComputer assisted language learningsConfidence measuresGoodness of Pronunciation (GOP)Goodness of pronunciationsLDA classifiersPronunciation error detectionPronunciation error detectionsSpeech communicationspeech
TNO Identifier
19222
Source title
Interspeech 2007
Pages
1837 - 1840
Files
To receive the publication files, please send an e-mail request to TNO Repository.