Applying the J48 decision tree classification algorithm in emotion recognition

Abstract

Physiological signals are external expressions of emotions. Mood changes can be expressed by changes in physiological signals. Because these performances are not controlled by the individual's subjective consciousness, the conclusions will be more objective and correct. The method based on statistical features is difficult to describe the complex changes of physiological signals, so the J48 decision tree is used to train and identify the chaotic characteristics of physiological signals in article. It has many advantages in solving multi-class or classification problems, such as high accuracy, fast classification speed, and simple classification rules. The chaotic feature matrices include the extracted chaotic feature parameters, which are combined with the J48 decision tree classifier to recognize four different emotions. The results show that emotion recognition of physiological signals based on chaos theory is feasible.