An Intelligent Cognitive-Inspired Computing for Complain Speech Analysis and Classification

This study trained and tested 1006 words using an LSTM network and a SOC model. The model achieved 90% accuracy and yielded relatively successful SOC corpus and SNA results. The computer model was also integrated into speech act theory by combining linguistics, computer science, and psychology. A large number of complaint languages were analyzed. This study demonstrates the feasibility of computer modeling and SOC in cognitive computing, showing their promising applications.