eBoF: Interactive Temporal Correlation Analysis for Ensemble Data Based on Bag-of-Features

Abstract

We propose eBoF, a novel time-varying ensemble data visualization approach based on bag-of-features (BoF). In the eBoF model, we take a simple and monotone interval from all target variables of ensemble scalar data as a local feature patch of BoF model and the duration time of each interval (i.e., feature patch) as its frequency. The feature clusters in ensemble runs are then identified based on the similarity of temporal correlations. eBoF generates the clusters together with their probability distribution across all the feature patches while storing the geo-spatial information, which is often lost in the traditional topic modelling or clustering algorithms. The probability distribution across different clusters can help to generate reasonable clustering results evaluated by the domain knowledge. We conduct several case studies and performance analyses. We also consult the domain experts to evaluate the proposed eBoF model. Evaluation results suggest the proposed eBoF can provide insightful and comprehensive evidence on ensemble simulation data analysis.