Research on Identifying Emotion Sentence Based on Fuzzy Information Granulation

Abstract

We put forward a method based on fuzzy information granulation to solve the problem of identifying the emotional sentence. Firstly, the necessary work is carried out on the microblogging sentence training set, including the word segmentation, stop words and so on. The feature vector is constructed and the weight of the feature is calculated. The fuzzy granulation algorithm proposed by W. Pedrycz is improved, and finally the improved the fuzzy information granulation algorithm learns the classification model and verifies the performance of the classification model on the test set. The experiment shows that this method has good recognition performance.