Network Time Series
2015 Joint Statistical Meetings, Society and Networks
In the modern era of science and technology, people are surrounded by, as well as participating, all different kinds of networks. One renowned example is social network or communication network. In general, networks can be classified into two large categories: static network and temporal network. Static network concerns about the network at a specific time point while temporal network focuses on the evolution or time series of networks.
Our research focus on temporal networks. We intend to develop tools incorporating graph network theories and time series analysis. Essentially, we are trying to investigate the temporal network data from two angles: node and link analysis. Node is defined as the member in the temporal network and link is defined as the interaction times between nodes. In node analysis, we want to find communities dynamically and also influential nodes. In link analysis, we want to understand and forecast the link time series by cluster analysis, change point analysis and multivariate time series analysis. In the meantime, we would like to propose methods to detect anomaly in network time series, which involves designing new statistics and making inference.
Our research focus on temporal networks. We intend to develop tools incorporating graph network theories and time series analysis. Essentially, we are trying to investigate the temporal network data from two angles: node and link analysis. Node is defined as the member in the temporal network and link is defined as the interaction times between nodes. In node analysis, we want to find communities dynamically and also influential nodes. In link analysis, we want to understand and forecast the link time series by cluster analysis, change point analysis and multivariate time series analysis. In the meantime, we would like to propose methods to detect anomaly in network time series, which involves designing new statistics and making inference.