A supervised approach to time scale detection in dynamic networks

• DOI: 10.48550/arxiv.1702.07752
• Publication Date: 2017-01-01
• Authors: Benjamin Fish; Rajmonda S. Caceres
• Keywords: Leverage (statistics); Granularity

For any stream of time-stamped edges that form a dynamic network, an important choice is the aggregation granularity analyst uses to bin data. Picking such windowing data often done by hand, or left up technology collecting However, can make big difference in properties network. This time scale detection problem. In previous work, this problem solved with heuristic as unsupervised task. As problem, it difficult measure how well given algorithm performs. addition, we show quality dependent on which task wants perform network after windowing. Therefore should not be handled independently from rest analysis We introduce framework tackles both these issues: By measuring performance based accomplished resulting are for first able directly compare different algorithms each other. Using framework, take supervised approach: they leverage ground truth training find good test approach approaches and several baselines real