About
CUMA is a Culture-aware User Modeling and Analysis framework that allows us to analyze and compare microblogging behavior for users from different cultural groups. It features techniques including semantic analysis and sentiment analysis that allows us to investigate not only the meaning of the microposts but also the users' opinions revealed in both Chinese and English microposts. Furthermore, we apply temporal analysis to analyze and compare how users' microblgging behavior changes over time on Sina Weibo and Twitter. Using these techniques, we conducted a data-intensive analysis based on Chinese (Sina Weibo) and English (Twitter) microblogging data to compare users' microblogging behavior between Chinese users and Western (American) users. We relate our findings to theories about cultural stereotypes developed in social sciences and therefore explain how our insights can allow for culture-aware user modeling.
Objective
We have given an innovative basis for analyzing microblogging behavior for users from different cultural groups. In our extensive studies (using 46 million messages from Twitter and Sina Weibo), we have highlighted the key differences in Western/US and Chinese practices, which partially reflecting underlying cultural differences. We have also explored the correlation between some of these differences and cultural models from social science research. Further interpretation and validation of our first set of conclusions can be done in future work, with research questions that follow our conclusions. Independent from these interpretations, our findings already provide valuable insights for the application using user modeling techniques that are provided by our user modeling framework.
Example scenarios
In the ImREAL project, our culture-aware user modeling framework is used to create a cultural user profile based on usage data collected from Social Web. Cultural profiles are utilized in a number of use-cases. For example, in the medical use case, which focuses on training medial students to conduct doctor-patient interviews. A cultural user profile can give an indication to the simulator about the user's culturality with a particular country, which in turn can be used inside the simulation to adapt the feedback the learner receives.
Services
Our culture-aware User Modeling framework features services that allows for conducting data-intensive analysis based on microblogging data and comparing users' microblogging behavior between Chinese users and American users, including:
- Syntactic analysis that extracts syntactical characteristics from microblogging data.
- Semantic analysis and sentiment analysis that investigate not only the meaning of the microposts but also the users' opinions revealed in both Chinese and English microposts.
- Temporal analysis that analyzes and compares how users' microblogging behavior changes over time on microblogging platform.
For ImREAL the ImREAL project, we have implemented the sentiment analysis as a Web service:
- sentimentAnalysis: The service takes a string as input and returns the sentiment, either positive, negative or neutral.
- TwitterSentiments: The service collects the user's past 200 tweets, filters out retweets and determines for each tweet the conveyed emotion (positive, neutral, negative). An aggregate score (between -1 and 1) is also computed which is an estimate of the overall emotion: a value close to 1 is heavily
positive, close to 0 means mostly neutral and -1 indicates very negative sentiments.
Datasets
We have collected more than 22 million posts from Sina Weibo and 24 million posts from Twitter. The dataset is available upon request. Contact Qi Gao for more details.
Publications
- Qi Gao, Fabian Abel, Geert-Jan Houben, Yong Yu. Information Propagation Cultures on Sina Weibo and Twitter. In Proceedings of ACM Conference on Web Science (WebSci), Evanston, USA, 2012 [bib, pdf]
- Qi Gao, Fabian Abel, Geert-Jan Houben, Yong Yu. A Comparative Study of Users' Microblogging Behavior on Sina Weibo and Twitter. In Proceedings of International Conference on User Modeling and Personalization (UMAP), Montr