Emotional Tendency Identification for Micro-blog Topics Based on Multiple Characteristics

Copyright 2012 by Quanchao Liu, Chong Feng, and Heyan Huang 26th Pacific Asia Conference on Language,Information and Computation pages 280–288 Emotional Tendency Identification for Micro-blog Topics Based on Multiple Characteristics Quanchao Liu Chong Feng Heyan Huang Department of Computer Science and Technology Beijing Institute of Technology Beijing, China {liuquanchao, fengchong, hhy63} @bit.edu.cn Abstract Public opinion analysis for micro-blog post is a new trend, and wherein emotional tendency analysis on micro-blog topic is a hot spot in the sentiment analysis. According to the characteristics of contents and the various relations of Chinese micro-blog post, we construct the dictionaries of sentiment words, internet slang and emoticons respectively, and then implement the sentiment analysis algorithms based on phrase path and the multiple characteristics for emotional tendency of micro-blog topics. Using micro-blogs’ forwarding, commentaries, sharing and so on, We take a future step to optimize the algorithm based on the multiple characteristics. According to the experimental results, our approach greatly improves the performance of emotional tendency identification on micro-blog topic. 1 Introduction As the fast development of New media, the internet gradually advocates open architecture philosophy that users participate in actively, and has been developed from a simple “reading webpage” to “writing webpage”, “building webpage together” for users. So a huge user generated content (UGC) has been developed, especially the emergence of micro-blog post. According to the reports1, users applying micro- 1 The Eleventh China Network Media Forum blog have exceeded 300,000,000. As a result of releasing diversely and writing randomly, micro- blog post is more and more popular in China, and it has brought a tremendous effect on network opinions and human society. Micro-blog contents formed on internet express people’s various emotion and sentiment, such as joy, anger, grief, praise, criticism and so on. More and more people like to use micro-blog post to share their views or experience, which makes people’s opinion information expanded rapidly. So it is very difficult to rely on the artificial method to deal with the micro-blog information’s collection and processing, there is an urgent need to help user analyse the massive information using computer. Chinese micro-blog post has multiple characteristics as following: 1) Due to the short text message, micro-blog post has terms’ sparsity. Feature extraction based on terms is not suitable for micro-blog post. 2) There exist many homophonic words, abbreviated words, internet slang in micro- blog post, such as “杯具” standing for “悲 剧 ”, “ 亲 ” standing for “ 亲 爱 的 ”, “3Q” standing for “谢谢”, or using emoticons to express emotion and so on. 3) Many popular new words made by network events would appear in micro-blog post. Taking “皮鞋很忙” for example, it appeared because the news report that the materials for producing shoes are processed into edible gelatin. 4) There are a variety of relations between micro-blogs. It is very convenient to forward, comment and share micro-blog post. 280 According to the above characteristics of micro-blog post, we analyse emotional tendency on micro-blog topic, and obtain a set of feasible …

Content truncated to 500 words