The turn up for this event was so good that the seminar room was packed with students, guests and distinguished speakers!
It was indeed an insightful workshop that shared current research works, and also showcased how the area of data mining in user generated content is going to be a topic so exciting! Speakers spoke with much enthusiasm with what they have to share and are optimistic on how these works can be useful and practical in organsations.
This workshop had a balance of both theories and demonstrations of applications. We had distinguished guest speakers from universities and from SMU too! Research works presented can be drawn back to some of the basic concepts learnt in one of our modules – Data Mining. (^_^)
Professor Bing Liu from University of Illinois at Chicago to share his research work on a technique that is able to mine opinions, by capturing the target object, the feature to be analysed, the opinion and time. He has suggested a way of creating a data structure to organise data such that opinion mining can be conducted. His concept of Divide and Conquer proofs convincing and workable, no wonder he is so optimistic of his work.
Professor Aixin Sun covered the topic using the approach of existing platforms like Wikipedia, Yahoo! Answers and Epinions. Professor Jing Jing also covered the area of domain adaptive information extraction in situations where training data is not available for information extraction. Professor Ee-Peng Lim also shared his study on Ratings intelligence using web 2.0. and how his idea of bia-ness and leniency affecting user ratings is able to discover the user’s behaviors.
Audiences had the privilege to witness how opinion mining works in a real application. We had a budding start-up company JamiQ (http://jamiq.com/; http://blog.jamiq.com/), to showcase the company’s product that provides opinion mining and sentiment analysis service. The application is able to understanding the sentiments of a sentence based on the score it generates. For e.g, “I like the user-friendliness of the iPhone.” will generate a positive score of “0.8″. This application has shown its potentials and how it can be a powerful tool in the area of marketing where businesses wants to understand the opinions of its potential consumers.
Another interesting research is by Data Mining Department at Institute for Infocomm Research, where they showed how data mining is being used to identify rising stars in organizations. In their research context where the bibliography networks is being mined, their algorithm uses the concept of co-authorship relationship and creating a human social network. A live demonstation was able to pull out the research works of a person in a database and identify if he is a potential hire in the company. It was really an amazing application that comes with visualisation that really aids decision making, and the potential such technology in organisations.
Well, this is indeed another insighful seminar that was not to be missed!
The recorded video of the seminar can be viewed at: http://media.smu.edu.sg/sistalk.htm
0 Responses to “Mining User-Generated Content Workshop”