In real-life scenarios, artificial customer service is capable of accurately recognizing customer feedback emotions. However, it faces limitations in terms of working hours and high costs. While intelligent customer service can handle most tasks, it lacks the ability to perceive real-time changes in user emotions through tone, thus affecting the overall user experience. Considering these social pain points, this paper proposes the use of Support Vector Machine (SVM) algorithm for Speech Emotion Recognition. By extracting ten distinct speech features and applying classification techniques, the results indicate that the model achieves an accuracy rate of 90%. The system consists of both hardware and software components. When the system detects user input data, it autonomously analyzes the emotions and generates a radar chart as the result. Human and machine customer service representatives utilize the emotion recognition results to develop a tailored solution for the user. A mobile application displays the detection results and provides specific solutions based on different emotions. In addition to its application in intelligent customer service and other smart interactive domains, this system can also be applied in various fields such as healthcare and transportation safety.
Paper
Full text
Intelligent Customer Service System Based on Speech Emotion Recognition
Semantic Scholar · Computer Science · 2023
Abstract
In real-life scenarios, artificial customer service is capable of accurately recognizing customer feedback emotions. However, it faces limitations in terms of working hours and high costs. While intelligent customer service can handle most tasks, it lacks the ability to perceive real-time changes in user emotions through tone, thus affecting the overall user experience. Considering these social pain points, this paper proposes the use of Support Vector Machine (SVM) algorithm for Speech Emotion Recognition. By extracting ten distinct speech features and applying classification techniques, the results indicate that the model achieves an accuracy rate of 90%. The system consists of both hardware and software components. When the system detects user input data, it autonomously analyzes the emotions and generates a radar chart as the result. Human and machine customer service representatives utilize the emotion recognition results to develop a tailored solution for the user. A mobile application displays the detection results and provides specific solutions based on different emotions. In addition to its application in intelligent customer service and other smart interactive domains, this system can also be applied in various fields such as healthcare and transportation safety.