Research On Emotion Detection From Texts

With the growing prosperity of Web 2.0,people are more and more inclined to share their feelings,attitudes and opinions through online news websites,blogs and other social network platforms,so that a large number of content generated by online users can be created.Analyzing the user-generated content(UGC)and detecting emotions from these texts can enhance the understanding of users' emotional state.Therefore,it is crucial to analyze and predict emotions from text accurately and to make computer to feel,understand and recognize human emotions.The main task of emotion detection from text is to detect the fine-grained emotions contained in the text or evoked by the text.Emotion detection has been applied in many downstream applications such as human-computer dialogue,recommendation system and public opinion monitoring.Therefore,the research of text-based emotion detection technology has important practical significance and application value.Most of the existing emotion detection methods only predict single emotion of the text,while the text in real world usually contains or stimulates a variety of different emotions of the public and these emotions have different emotional intensity.Therefore,it is very important to predict and rank multiple relevant emotions in the text.So this paper focuses on Relevant Emotion Ranking(RER)learning.The main contributions of this paper are as follows:1.Text might contain or invoke multiple emotions with varying intensities.As such,emotion detection,to predict multiple emotions associated with a given text,can be cast into a multi-label classification problem.We would like to go one step further so that a ranked list of relevant emotions are generated where top ranked emotions are more intensely associated with text compared to lower ranked emotions,whereas the rankings of irrelevant emotions are not important.A novel framework of relevant emotion ranking is proposed to tackle the problem.In the framework,the objective loss function is designed elaborately so that both emotion prediction and rankings of only relevant emotions can be achieved.Moreover,we observe that some emotions co-occur more often while other emotions rarely co-exist.Such information is incorporated into the framework as constraints to improve the accuracy of emotion detection.Experimental results on two real-world corpora show that the proposed framework can effectively deal with emotion detection and performs remarkably better than the state-of-the-art emotion detection approaches and multi-label learning methods.2.As emotions might be evoked by hidden topics,it is important to unveil and incorporate such topical information to understand how the emotions are evoked.We proposed a novel interpretable neural network approach for relevant emotion ranking.Specifically,motivated by transfer learning,the neural network is initialized to make the hidden layer approximate the behavior of topic models.Moreover,a novel error function is defined to optimize the whole neural network for relevant emotion ranking.Experimental results on three real-world corpora show that the proposed approach performs remarkably better than the state-of-the-art emotion detection approaches,RER and multi-label learning methods.Moreover,the extracted emotion-associated topic words indeed represent emotion-evoking events and are in line with our common-sense knowledge.3.Existing studies are based on either shallow representations such as bag of words without considering word ordering,or deep semantic representations using recurrent neural networks which have a difficulty in capturing long-distance dependencies.In this paper,we propose a novel hierarchical state recurrent neural network for relevant emotion ranking.Instead of incrementally reading a sequence of words as in traditional recurrent neural networks,the hierarchical state recurrent neural network encodes the hidden states of all words or sentences simultaneously at each recurrent step to better capture long-range dependencies.Moreover,the hierarchy mechanism is employed to capture the key hierarchical semantic structure in a document.Experimental results on two real-world corpora show that the proposed approach performs remarkably better than the state-of-the-art emotion detection approaches,INN-RER and multi-label learning methods.Furthermore,the hierarchy mechanism is employed to capture the key hierarchical semantic information of a document,which enables dynamically highlighting important parts in text evoking the emotions.