Call for Papers

Text information in the world can be roughly categorized into two main types: facts and opinions. Much effort has been invested in the past decades in the fact–based text processing, and many techniques have been developed for retrieving information or text mining. The opinion–based text processing, however, has received just limited attentions from researchers. Listening to other people´s opinions is important, especially when taking account of valuable customers´ opinions into an organization´s decision making process. However, identifying relevant sources, extracting topic features, and summarizing opinions are all difficult tasks. These are the problems targeted by topic feature discovery and opinion mining.

Topic feature discovery aims to identify on–topic sources and extract relevance features for a given topic (e.g., a person, an event, or a government policy). Many empirical experiments illustrated that the performance of text mining was hindered because of the measures shipped from data mining. Such performance measures, for example, support and confidence, turn out to be unsuitable in the leveraging stage. By way of illustration, given a specified topic, usually a highly frequent pattern (normally a short pattern) is general and a specific pattern is lowly frequent. The objective of topic feature discovery is to find a suitable subset of features available in text documents to describe the requested topic.

Opinion mining, also known as sentiment analysis, aims to summarize and classify opinions. Compared with traditional topic–based or fact–based analysis, opinion mining tends to address the new problems raised by the applications concerning the subjective or opinionated expression. Its aim is to determine the inclination of a reporter with respect to some topics, or to extract the opinions from a large variety of digital texts containing opinionated content. Difficult problems in opinion mining include extracting opinionated information; classifying sentiments and subjectivity; analysing feature–based sentiments; identifying opinion spam; and applying opinions to problem solving or decision making. Therefore, there are many opportunities and feasibility of extensive research activities in the field of opinion mining.

Topic feature discovery and opinion mining are extremely challenging topics in modern information analysis, from both an empirical and a theoretical perspective. They are also of central interest and the critical steps for many Web personalized applications and recommender systems. The problems in topic feature discovery and opinion mining have charged continuously increasing attentions from researchers in data mining, Web intelligence, text mining, machine learning, natural language processing, and information retrieval communities. Strongly focusing on these two challenging topics and their surrounding areas, this workshop aims to enhance the current text mining and opinion mining techniques, and explores a new methodology to discover and interpret useful and interesting knowledge in text documents.

Topics include, but are not limited to:

• Relevance feature discovery
• Opinion mining and sentiment analysis
• Information filtering and retrieval
• Text mining
• Text categorizations
• Ontology mining
• Information extraction
• Sentiment and subjectivity classification
• Feature–based sentiment analysis
• Recommender systems
• Web personalization
• Evaluation methodologies in topic feature discovery and opinion mining