Research
Social Media Analysis
Graph Mining and Human Factor Analysis in Social Media
Examining the Relationship between Profile Images and User Actions in Twitter
Workshop on Modeling Social Media 2015 (MSM 2015 c/w ACM WWW’15)
Researchers: Tomu Tominaga
This study examines the relationship between prole images and user behaviors on Twitter. We assume that the prole images and behaviors of users are influenced by their internal properties, because users consider their prole images as symbolic representations of themselves on Twitter. We empirically categorized prole images into 13 types, and investigated the relationships between each category of prole images and users’ behaviors on Twitter. [PDF]
Computing Semantic Relatedness using Layout Information of Wikipedia
Cognitive Informatics and Natural Intelligence, Vol.7, No.2, 2013.
Researchers: Patrick Chan
Computing the semantic relatedness between two words is an important problem in information retrieval and natural language processing. Explicit Semantic Analysis (ESA), a state-of-the-art approach to solve the problem represents a word as concept vector using the article of Wikipedia. The value is calculated using td-idf. To improve the estimation accuracy, we use not only the word’s frequency, its location in an article, and its text style. Empirical evaluation shows that, for low-frequency words, the our method achieves better estimate of semantic relatedness over ESA. [PDF]
Community Extracting using Intersection Graph and Content Analysis in Complex Network
IEEE/WIC/ACM International Conference on Web Intelligence’12
Researchers: Toshiya Kuramochi, Naoki Okada, Kyohei Tanikawa
One hot topic in the study of complex network (social network) is community detection. This research proposes a new method for finding communities in complex networks. Our proposed method considers the overlaps between communities using the concept of the intersection graph. Additionally, we address the problem of edge in homogeneity by weighting edges using the degree of overlaps and the similarity of content information between sets. Finally, we conduct clustering based on modularity. Our method ourperforms the conventional method when applying it to a real SNS network. [PDF]