This project develops methods and models for understanding and predicting attention dynamics across online platforms. The project consists of five successive topics focusing on measurements, data sampling, and predict...
详细信息
This project develops methods and models for understanding and predicting attention dynamics across online platforms. The project consists of five successive topics focusing on measurements, data sampling, and predictive models for social processes. First, we showideological asymmetries in digital space and provide a set of methods to quantify attention dynamics across different social platforms, especially YouTube and Twitter on long-running controversial topics. Second, we measure the correlation between online behavior and offline attitudes and actions, grounded on the theory of discursive opportunities. Third, we present a first study on cross-partisan communications on YouTube comments and find that the crosstalk is not symmetric. Fourth, we present a first study on measurement errors under subsampled Twitter data streams, and discuss noises and potential biases in social data. Lastly, we develop three different models that explain how social process unfolds: a mathematical relationship between self-exciting processes and stochastic epidemic models; a succinct neural model that universally approximates any point process; and a dual mixture model that is particularly suited to long-tailed data with both popular and unpopular content.
The theory of wavelets has recently undergone a period of rapid development. We introduce a software package called wavethresh that works within the statistical language S to perform one- and two-dimensional discrete ...
详细信息
暂无评论