A1257
Title: Scanner: Simultaneously temporal trend and spatial cluster detection for spatial-temporal data
Authors: Xin Wang - San Diego State University (United States) [presenting]
Abstract: Identifying the underlying trajectory pattern in the spatial-temporal data analysis is a fundamental but challenging task. The problem of simultaneously identifying temporal trends and spatial clusters of spatial-temporal trajectories is studied. To achieve this goal, a novel method named spatial clustered and sparse nonparametric regression (scanner) is proposed. The method leverages the B-spline model to fit the temporal data and penalty terms on spline coefficients to reveal the underlying spatial-temporal patterns. The method estimates the model by solving a doubly-penalized least square problem, in which a group sparse penalty for trend detection and a spanning tree-based fusion penalty for spatial cluster recovery are used. An algorithm based on the alternating direction method of multipliers (ADMM) algorithm is developed to efficiently minimize the penalized least square loss. The statistical consistency properties of scanner estimator are established. Thorough numerical experiments are conducted to verify the theoretical findings and validate that the method outperforms the existing competitive approaches.
