Application of a hybrid speech-recognition post-processing system to improve the quality of decision-making based on phonetic analysis of words and phonemes

• Published in: AIP conference proceedings Vol. 2849; no. 1
• Main Authors: Tarasiev, Andrey A., Filippova, Margarita E., Aksyonov, Konstantin A., Talantsev, Egor N.
• Format: Journal Article Conference Proceeding
• Language: English
• Published: Melville American Institute of Physics 01.09.2023
• Subjects: Algorithms; Decision analysis; Hybrid systems; Phonemes; Speech recognition
• ISSN: 0094-243X; 1551-7616
• DOI: 10.1063/5.0162428

In the article, we consider the idea of developing an intelligent and customizable automated system for maintaining text and voice dialogs with the user in the real time. This system can be used for almost any subject area, for example, to create an automated robot - the operator of the call center or smart chat bots, assistants and so on. The work in details presents module of recognized text phonetic post-processing on the example of the Russian languages. This is improvement of existing post-processing submodule used application of different Stemming methods and stochastic keyword ranking using weights table. The phonetic analyze submodule uses word transformations and analyze of phoneme similarity using Levenshtein distance algorithm.