Statistical Analysis of Voice Based Emotion Recognition using Similarity Measures

Emotion recognition is gaining more and more importance. Emotional AI systems are allowed to detect, analyze, process and respond to people's emotional states and moods. One specific application could include car navigation systems that are able to hear a driver start to experience road rage, and react to prevent them from making a rash driving decision. Another similar one could be used to allow automated assistants to change their approach when they hear anger or frustration from a user. Present voice assistance provides information to the user's speech. This paper focuses on categorizing the emotions and applying statistical measures to identify the similarities among the different features. Preprocessing and Feature recognition is carried out using R statistical tool. The results are analyzed by identifying the similarities using Jaccard, Cosine and correlation similarity measures.