Big data analysis voice recognition method

• Main Authors: MA RUYAN, LI LEI
• Format: Patent
• Language: Chinese; English
• Published: 09.10.2020
• Subjects: ACOUSTICS; MUSICAL INSTRUMENTS; PHYSICS; SPEECH ANALYSIS OR SYNTHESIS; SPEECH OR AUDIO CODING OR DECODING; SPEECH OR VOICE PROCESSING; SPEECH RECOGNITION

The invention discloses a big data analysis voice recognition method. The big data analysis voice recognition method comprises the steps that: A, after voice signals to be recognized are collected, the voice signals to be recognized are transmitted to a classification module, the classification module divides the voice signals to be recognized into simple vocabulary signals and complex vocabularysignals, the simple vocabulary signals are transmitted to a first recognition module, and the complex vocabulary signals are transmitted to a second recognition module; B, the first recognition modulecompares the simple vocabulary signals with data in a database to obtain a preliminary screening data set of which the similarity with the simple vocabulary signals is higher than a set threshold value; and C, the second recognition module determines recognition parameters according to the preliminary screening data set, and after the second recognition module recognizes the complex vocabulary signal, data with the maximu