| The rapid development of the Internet, making the Internet a life of important people, people Wangluo published a large number of comments, messages and so on. Through these valuable comments of the network, people can get information about customer attitudes. Allows people to comment on the emotions of the network had a strong interest. At present, the emotional analysis of the English have been doing a lot of research and achieved great results. In contrast Chinese sentiment mining has done research related to small but also fragmented.The purpose of this paper is in English under the PMI-IR model and the Chinese environment A PMI-IR model based on the proposed review of a more emotional adaptation of Chinese mining model B. And B in this model based on the movie reviews of establishing emotional for network mining model.This contribution includes the following aspects:1. In this paper, the PMI-IR model of the English and Chinese PMI-IR model based on A, put forward even more to adapt to the Chinese environment, PMI-IR model B. Model B key word in the phrase in reference to the choice of mode and made the appropriate improvements, and apply online movie review mining of experimental verification of the model B of the advanced nature. Model B manual aspects of accuracy and made great improvements, improved accuracy and reduced manual workload.2. Movie reviews on the web mining process in the precision and recall rate of the phenomenon of low-error analysis. Model B in the application of the results of emotion mining, to identify the reasons for the lower results, and discussed analysis.3. B on the basis of the model to explore the establishment of a network of movie reviews for the mining model for emotional tendencies. And apply this model to the preliminary experiment has shown good results.4. This article reviews the signs on the network conducted a preliminary study of emotion, and the corresponding emotional excavation analysis algorithm. Because the network, the people are now accustomed to a variety of symbols to express their feelings, and our comments in order to dig Chu Qing Xiang Ping Lun those feelings, the need for symbols to be on the Zhexieqinggan excavation.The whole article is to discuss the following ideas:This paper describes the research background and current English and Chinese analysis of the emotional status, at present, English-based excavation analysis of the emotional environment more, and has a high accuracy and for feelings of the Chinese networks are less powerful, who leaves the first time, PMI-IR model into Chinese, and the corresponding improvement has been successful. In this context, this paper, the current sentiment analysis in English and Chinese method of feeling, then proposed a new PMI-IR model B. The model B in the extract and the reference word phrase pattern corresponding changes done in the phrase pattern extraction, the article reviews characteristics of the network, statistical model proposed six phrases; in the reference terms, this paper comments from the network to automatically statistical methods, rather than manually determine aspects. Next, in accordance with the PMI-IR model B, the film comments on the network excavation analysis, and finally through the analysis of experimental results verify the feasibility of the model B. And the application of this model according to the results of experiment B, the discussion of the movie review mining for network analysis of the reasons for the lower precision. Based on the model B, the paper discusses the network of movie reviews for the mining model of emotion. Mining in the network model of emotional movie reviews, text, symbolic expression of emotion for the first time digging through the assignment of the expression analysis of symbols, and then be judged by determining the threshold, can solve the problem emoticons, get better The mining results.Text mining model to determine the Chinese emotion PMI-IR model B, application of semantic analysis for future research to identify the emotional tendency of a more advanced model, to facilitate future research on the emotional tendencies. This paper presents the network model movie review mining is a movie review for the network analysis model of emotion, the film movie review site by the excavation of network analysis can determine the emotions contained in them tend to determine what decision-making, success.
|
PDF Full Text Request