High Influencing Pattern Discovery over Time Series Data

Dianwu Fang,Meijiao Wang,Lizhen Wang,Jialong Wang

doi:10.3390/ijgi10100696

Abstract

A spatial co-location pattern denotes a subset of spatial features whose instances frequently appear nearby. High influence co-location pattern mining is used to find co-location patterns with high influence in specific aspects. Studies of such pattern mining usually rely on spatial distance for measuring nearness between instances, a method that cannot be applied to an influence propagation process concluded from epidemic dispersal scenarios. To discover meaningful patterns by using fruitful results in this field, we extend existing approaches and propose a mining framework. We first defined a new concept of proximity to depict semantic nearness between instances of distinct features, thus applying a star-shaped materialized model to mine influencing patterns. Then, we designed attribute descriptors to perceive attributes of instances and edges from time series data, and we calculated the attribute weights via an analytic hierarchy process, thereby computing the influence between instances and the influence of features in influencing patterns. Next, we constructed influencing metrics and set a threshold to discover high influencing patterns. Since the metrics do not satisfy the downward closure property, we propose two improved algorithms to boost efficiency. Extensive experiments conducted on real and synthetic datasets verified the effectiveness, efficiency, and scalability of our method.

Highlights

Construct new proximate relations: In order to construct new proximate relations between instances of distinct features, we reviewed the theoretical basis by which traditional spatial co-location pattern mining adopts spatial proximity, i.e., the first law of geography (or called Tobler’s first law (TFL)), alleging that everything is related to everything else but near things are more related to each other [8,9]
For an influence propagation process abstracted from epidemic dispersal scenarios, we define a semantic proximate relationship to describe the nearness between instances of distinct features and propose a novel influencing pattern, and we further introduce a high influencing pattern based on the directions of influence and inner changes in instances perceived by an attribute-aware analysis of time series data to meet the needs for pattern discovery in the influence propagation process
To aid the intuitive understanding of the influence propagation process proposed in this study, we provide a general overview in Figure 1, where an epidemic first outbreaks in instance B.1 and spreads along influential media flows within a period of time

Summary

Introduction

Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. In the past two decades, spatial co-location pattern mining has been a hot topic in the field of spatial data mining. After Shekhar et al [1] first introduced the notion of spatial co-location patterns in 2001, many experts and scholars devoted themselves to this field and achieved abundant results. The mining of spatial co-location patterns and extended patterns has been widely applied in public governance and traffic management and services, among others [2,3,4]

Objectives

Methods

Conclusion