DocumentCode
2457014
Title
Automatic identification of positive or negative language
Author
Posner, Erez ; David, Omer ; Aharonson, Vered ; Shafat, Gabi
Author_Institution
Afeka, Tel Aviv Acad. Coll. of Eng., Tel Aviv, Israel
fYear
2012
fDate
14-17 Nov. 2012
Firstpage
1
Lastpage
4
Abstract
Personal coaching, performed by professionals such as psychologists, usually includes training for business as well as social situations such as job interviews, business meetings, interaction with a customer service provider, and more. This requires careful preparation in which, among other traits, the trainees need to pay attention to the words they choose in the interaction, in order to make a positive impression. To achieve this goal, we have developed a coaching system using speech recognition, which enables both monitoring by the coaching professional and self-training by the user. By providing timely indications as to when the user employs positive or negative expressions as defined by the psychologist, the system helps users develop self-control and awareness regarding the language they use. The system consists of adjusted voice activity detection (VAD) and key word spotting (KWS) algorithms, implemented together with an interactive UI into an Android-based application, available on cellular phones.
Keywords
mobile handsets; natural language processing; speech recognition; user interfaces; Android-based application; KWS algorithms; VAD; adjusted voice activity detection; cellular phones; interactive user interface; keyword spotting algorithms; negative language automatic identification; personal coaching system; positive language automatic identification; speech recognition; Feature extraction; Real-time systems; Smart phones; Speech; Speech recognition; Training; Vectors; KWS; VAD; android OS development; signal processing; speech recognition;
fLanguage
English
Publisher
ieee
Conference_Titel
Electrical & Electronics Engineers in Israel (IEEEI), 2012 IEEE 27th Convention of
Conference_Location
Eilat
Print_ISBN
978-1-4673-4682-5
Type
conf
DOI
10.1109/EEEI.2012.6377047
Filename
6377047
Link To Document