1461174628-83c1d4f0-2573-4f61-95fb-95e7d5cda73c

A speech recognition method and apparatus is disclosed in which each utterance matched to a feature model within a feature model memory is associated with a confidence score indicative of the posterior probability of the word being correctly matched, given that the matching of the utterance to the feature model generated certain values indicative of the goodness of the match. The method may include identifying a first operand and a second operand. There also is a movable shade rail at the end of the second shade material. The phone therefore alerts the user to the incoming call not noticed or answered by the user despite that the user carries it.