TIMan Logo Text Information Management Group
Home | About| People | Projects| Publications | Downloads| Links| Wiki


Existing work on leveraging opinionated content has been focused on generating structured summaries of opinions to help users better digest all the opinions. Unfortunately, in many decision making scenarios, ranging from what product to purchase to which doctor to consult for a particular ailment, the choices are often a plenty and summaries alone will not suffice in helping users reach their decision.

We propose a different way of leveraging opinions, that is to combine the power of search technologies with opinion analysis and mining tools to provide a unified decision making platform. We call this synergistic platform an Opinion-Driven Decision Support System (ODSS) - a system that enables users to find and analyze entities of interest (e.g. products, people and businesses) based on the opinions of other web users. Our research encompasses (1) Searching capabilities based on opinions, (2) Different forms of opinion analysis tools, (3) Opinionated data collection and (4) Presentation of results.



view screenshot FindiLike is preference driven search engine that enables users to find entities based on personal preferences. These preferences can be a combination of unstructured information such as preferences for opinions or structured ones such as price, distance and so on depending on the domain. Currently it provides support for hotel search. The opinion preferences can be expressed as a set of natural keywords such as 'close to universal studios', 'safe location', 'good breakfast', etc.
[ More Info | System ]

Selected Publications

FindiLike: Preference Driven Entity Search, Kavita Ganesan and ChengXiang Zhai, Proceedings of the 21st International Conference on World Wide Web 2012 (WWW '12), Demo

Micropinion Generation: An Unsupervised Approach to Generating Ultra-Concise Summaries of Opinions, Kavita Ganesan, ChengXiang Zhai and Evelyne Viegas,Proceedings of the 21st International Conference on World Wide Web 2012 (WWW '12)

Opinion-Based Entity Ranking, Kavita Ganesan and ChengXiang Zhai, Information Retrieval, vol. 15, issue 2, 2012

Opinosis: A Graph Based Approach to Abstractive Summarization of Highly Redundant Opinions, Kavita Ganesan, ChengXiang Zhai and Jiawei Han, Proceedings of the 23rd International Conference on Computational Linguistics (COLING '10)


For more information about this project please contact Kavita Ganesan or ChengXiang Zhai