Call for Papers – Social Media Mining

For example, consider the 13″ Retina MacBook Pro. At the time of writing its features, as listed by Amazon, cover technical details such as screen-size, RAM, processor speed, and price. These are the type of catalog features that one might expect to find in a conventional content-based recommender. But often such features can be difficult to come by and they can be technical in nature, thereby limiting recommendation opportunities and making it difficult for casual shoppers to judge the relevance of suggestions in any practical sense. However, the MacBook Pro has more than 70 reviews that encode valuable insights into a great many of its features, from its beautiful design and great build quality to its high price. These features capture more detail than a handful of technical (catalog) features. They also encode the opinions of real users and, as such, provide an objective basis for product comparisons.

In this talk we will explore the use of opinion mining techniques to extract meaningful product descriptions from user-generated reviews and evaluate their use in a novel approach to product recommendation that combines product similarity and feature sentiment. We will describe the results of a recent evaluation using Amazon and TripAdvisor data to demonstrate the utility and generality of this approach for recommendation.

Session 1:

- New methods or approaches for mining social media data and new approaches to model users or other entities in social media – Chair: Dr Nirmalie Wiratunga.
Presentations:
• Adam Wyner: Semi-Automated Extraction of Arguments from Online Product Reviews.
• Sadiq Sani, Nirmalie Wiratunga, Stewart Massie and Robert Lothian (presented by Sadiq Sani): Sentiment Classification using Supervised Sub-Spacing

Session 2:

-Evaluation methods for mining and modelling in social media – Chair: Prof. Ayse Goker

Presentations:
• David Corney, Ayse Goker and Carlos Martin (presented by David Corney): Evaluating information retrieval in Social Media
• James McMinn: Issues in developing a test collection for evaluating event detection on Twitter
• Ben Horsburgh, Susan Craw and Stewart Massie (presented by Ben Horsburgh): Evaluating Music Recommendation Using Implicit and Explicit Feedback from Social Media

Social Media (especially Twitter) is widely seen as a source of real-time breaking news. For example, when Osama Bin Laden was killed by US forces the news was first made public on Twitter. Rapidly finding all breaking news has clear economic and humanitarian benefits.
Finding all such breaking news presents hard computational challenges. We need to detect news-related novelty in massive streams (upwards of two thousand posts per second) as quickly as possible. Efficiency is not the only consideration however and we also need to confront the enormous quantity of irrelevant posts. In this talk I will outline how we tackle the first problem using Locality Sensitive Hashing, taking constant time per post. In tandem I will discuss how we use Storm to parallelise this computation, yielding a system capable of processing 2k tweets per second. The second problem is tackled by intersecting the Twitter stream with Wikipedia page requests, filtering-out spurious first stories. Taken together, this results in processing more than 250 million items per day.
Joint work with Sasa Petrovic (Edinburgh), Craig MacDonald (Glasgow), Iadh Ounis (Glasgow) and Richard McCredie (Glasgow)

Session 3:

Applications and visualisations of social media mining – Chair: Dr. Carlos Martin- Analytics in social media.

Presentations:
• Elias de Oliveira, Patrick Marques Ciarelli, and Marco Toledo (presented by Elias Oliveira): Towards a Classification System for 1 Billion of Tweets
• Aminu Muhammad, Nirmalie Wiratunga, Robert Lothian and Richard Glassey (presented by Aminu Muhammad): Sentiment Retrieval and Visualization in Discussion
• David Wilson: Mining Crowdsourced Data for Geographic Knowledge