ISSN : 1796-203X
Volume : 2    Issue : 10    Date : December 2007

Sequence Package Analysis: A New Method for Intelligent Mining of Patient Dialog, Blogs and
Help-line Calls
Amy Neustein
Page(s): 45-51
Full Text:
PDF (288 KB)

The ambiguities, repetitions and ellipses commonly found in natural language dialog continue to
hinder speech (and text) analytic mining programs that glean business intelligence data from
consumer help-line calls, or extract important medical diagnostic information from doctor-patient
interviews or consumer-generated health-related blogs. This poses an even greater problem when
such mining programs attempt to extract critical emotional data from natural language dialog. At
present, speech (and text) analytic programs that mine natural language dialog for signs of
distress, frustration, anger or other human emotions are still largely ineffective, because
conventional speech systems that are limited to a set of key words and phrases cannot process
speech as it actually occurs; if a speaker or blogger fails to use the word(s) found in the speech
application’s vocabulary, the program yields a poor statistical word match (or no match). This paper  
shows how Sequence Package Analysis is informed by a set of algorithms – representing some of
the more complex semantic aspects of communication in addition to syntax – that can interpret less
than perfect natural speech, enhancing intelligent mining of recordings of doctor-patient interviews,
customer care help-line calls, and consumergenerated health-related blogs.

Index Terms
blogosphere, call centers, customer-care helpline calls, data mining, doctor-patient dialog, medical
interviews, natural language understanding, sequence package analysis, speech analytics