RETRACTION: Detecting anomalous emotion through big data from social networks based on a deep learning method (Retraction of Vol 19, Pg 251, 2018)

作者：Sun, Xiao^*; Zhang, Chen; Ding, Shuai; Quan, Changqin

来源：Multimedia Tools and Applications, 2020, 79(13-14): 9687-9687.

DOI：10.1007/s11042-018-5665-6

• Anomaly detection
• Convolutional neural network
• Gaussians
• Generalization ability
• Learning models
• Multivariate gaussian models
• Power law distribution
• Social media

Anomaly detection in social media refers to the detection of users’abnormal opinions, sentiment patterns, or special temporal aspects of such patterns. Social media platforms, such as Sina Weibo or Twitter, provide a Big-data platform for information retrieval, which include user feedbacks, opinions, and information on most issues. This paper proposes a hybrid neural network model called Convolutional Neural Network-Long-Short Term Memory(CNN-LSTM), we successfully applies the model to sentiment analysis on a microblog Big-data platform and obtains significant improvements that enhance the generalization ability. Based on the sentiment of a single post in Weibo, this study also adopted the multivariate Gaussian model and the power law distribution to analyze the users’emotion and detect abnormal emotion on microblog, the multivariate Gaussian method automatically captures the correlation between different features of the emotions and saves a certain amount of time through the batch calculation of the joint probability density of data sets. Through the measure of a joint probability density value and validation of the corpus from social network, anomaly detection accuracy of an individual user is 83.49% and that for a different month is 87.84%. The results of the distribution test show that individual user’s neutral, happy, and sad emotions obey the normal distribution but the surprised and angry emotions do not. In addition, the group-based emotions on microblogs obey the power law distribution but individual emotions do not.

• 出版日期2020-4
• 单位合肥工业大学