Speech recognition: Statistical methods

摘要: Statistical methods for speech processing refer to a general methodology in which knowledge about both signal and the language that it expresses, along with practical uses of specific tasks or services, is developed from actual realizations data through well-defined mathematical statistical formalism. For more than 20 years, this basic has produced many advances new results, particularly recognizing understanding natural by machine. In article, we focus on two important methods, one based primarily hidden Markov model formulation gained widespread acceptance as dominant technique characterizing variation acoustic representing speech, related use statistics word co-occurrences. This second acts form grammar set syntactical constraints language. contrast earlier systems employed linguistic analyses, these data-driven have proven produce consistent useful results become underpinning technology modern recognition systems. Such are used wide range applications such automatic telephone call routing information retrieval.