A Hybrid Approach to Automated Rating of Foreign Language Proficiency Using Oral Test Responses
This study was conducted to improve the automatic rating of oral test responses collected through Language Testing International’s (LTI) Oral Proficiency Interviews using a Computer (OPIc). In OPIc tests, a computer automatically asks questions from the candidate and the responses of the candidate are recorded and consequently rated. This study has been performed on English OPIc tests. Although, no specific knowledge of the English language has been used for this phase of research and the results may be readily extended to tests in other languages. Preliminary results are quite promising, considering the utilization of the crude Verbosity feature.
KeywordsLanguage Model Rating Level Human Rating Speech Recognizer Prompt Response
The author would like to thank the Center for Language Studies at the Brigham Young University, Language Testing International (LTI), and the American Council on the Teaching of Foreign Languages (ACTFL) for making this research possible.
- ACTFL (2012) American Council on the Teaching of Foreign Languages (ACTFL) guidelinesGoogle Scholar
- Beigi H (2008) Whether computer analyses can predict human ratings of speaking proficiency. Recognition Technologies, Inc. Technical Report No.: RTI-20081205-01Google Scholar
- Beigi H (2011) Fundamentals of speaker recognition. Springer, New York. ISBN: 978-0-387-77591-3, http://www.FundamentalsOfSpeakerRecognition.org
- Duda Richard O, Hart PE (1973) Pattern classification and scene analysis. Wiley, New York. ISBN: 0-471-22361-1Google Scholar
- G.711 (1988) Pulse Code Modulation (PCM) of voice frequencies. ITU-T recommendation. URL www.itu.int/rec/T-REC-G.711/e
- Viswanathan M, Beigi HS, Maali F (2000) Information access using speech, speaker and face recognition. In: IEEE international conference on multimedia and expo (ICME2000)Google Scholar