Yes/No Corpus
                            Release 1.2

              Center for Spoken Language Understanding


UPDATED: 23 August 2002


Corpus Description
------------------
This files describes some of the overall characteristics of the
corpus.  Some statistics concerning the number of each type of
utterance are given.

Sometimes a caller would answer a yes/no question with a response
other than a simple "yes" or "no".  For example, many people said "yes
there is" when we asked if there was an "a" in their last name.  We
eliminated these utterances from the corpus.  Likewise, when callers
said a word other than "yes" or "no" (e.g. "sure"), we eliminated
those files from the corpus. 

The corpus consists of about 7 hours of speech.

The following table gives a count of the number of files for each
utterance type.

	Type		Number
	----------------------
	cellular         270
	ever_married    4004
	have_home_phone 3923
	hispanic        3943
	letterA		1568
	no		2070
	results		1234
	yes		2040
	yorn		 476


The following table shows the total number of "yes" and "no" responses
in the corpus.

	Word		Number
	----------------------
	no		 8157
	yes		11371