A rough-set method generates prediction rules from observed data, usin
g statistical information inherent in the data to handle incomplete an
d ambiguous training samples. Experimental results indicate that this
method provides more precise information than is available through kno
wledge acquisition from human experts.