Speech Interaction’s Application In Database Information Retrieval

Speech interaction technology including speech recognition and text to speech becomes increasingly sophisticated, and has been applied to many advanced electronic devices in market. The combination of speech recognition technology with database information retrieval system to design the natural language interface with voice form facilitates the information access. Thus, it has great theoretical and practical value.This thesis aims to study natural language interface based on voice form used in information retrieval system. Firstly, the basic theory of speech recognition is introduced, including preprocessing, extraction of characteristic parameter and matching algorithm. Further, the advantages and disadvantages of man-machine interface based on speech recognition are discussed. It is proved the feasibility of the system according to the characteristics of Chinese query statement combined with relevant references. The design of knowledgebase is of great essential. Thus, the thesis focuses on exploring voice features knowledgebase in depth combined with the main task equivalent semantic conversion.With the theories and technologies above, a speech-form natural language interface is designed and implemented around the most common database instance student information retrieval system. It is noted that this system uses Visual Studio2008as the development platform, ADO technology as database communication part, Microsoft Access2007as the database and Speech SDK5.1published by Microsoft Ltd as speech recognition and text to speech module. In the final, I analyze the experimental results, discuss its deficiency and later continued improvements.