Machine learning techniques for document filtering
Richard M. Tong, Lee A. Appelbaum · 1993
Booz Allen & Hamilton's Advanced Decision Systems Division is conducting a program of research to investigate machine learning techniques that can automatically construct probabilistic structures from a training set of documents with respect to a single target filtering concept, or a set of related concepts. These structures can then be applied to individual documents to derive a posterior probability that the document is about a particular target concept.