Predicting State Transitions by Learning Spatio-Temporal Patterns
Investigator: Alex Sim
Newly developed analysis methods learn spatio-temporal relationship on streaming data, and reveal transitions in network operation states, offering a new way to classify and predict the anomalous state.
Significance and Impact
- Our research studied machine learning-based clustering method and density-based grid structure and joint distribution methods to understand variations in patterns for data with multiple features.
- Similarity measures have been designed to estimate temporal variations; "degree of change" based on moving distance of clusters and "common occupancy rate” (similarity) based on the concept of Jaccard Index for grid structure.
- Streaming data including network traffic monitoring measurements show the non-linear property, and the dimensionality reduction is a challenge. Deep learning models would be essential to find spatio-temporal variations in data streams with multiple features.