Dr. Pei-fen Kuo (郭佩棻)
Assistant Professor
Sub-Project Principal Investigator National Cheng Kung University |
Academic Biography
My research interest is human mobility and its interaction with the environment in space. Apart from data mining techniques, we also often use spatially-related methods including spatial statistical models, spatial analysis and trajectory analysis to understand social science hot issues such as human travel behavior, traffic safety, spatio-temporal patterns of crime or air pollution prevention strategies. We collect, analyze and interpret data of cities in Taiwan and overseas, and also have active international research participations and collaborations including conference attendance and presentations and international journal publications.
My recent research contributions can be summarized to two focuses: the human travel behavior and traffic safety. On human travel behavior, we analyzed the spatial effect of Americans with disabilities act (ADA) paratransit services in Houston and built up demand estimation models (travel demand models) for each zone. We used the geographical weight regression (GWR) model to capture spatial variations of each environmental factors, and proved to improve the predictive performance, compared with the ordinary least squares (OLS) regression model. The study was presented in international conference and published on the SCI journal, Transportation Research Record. We also did domestic research, analyzing point data of bike sharing services in Taipei City and Tainan City, and building up travel time estimation model from the origin-destination pairs. By using spatial models such as spatial survival analysis, this study explored the travel behavior and visualized dynamic network patterns of bike stops. The result was also presented in international conference and published on the SCI journal. On traffic safety, we collected the traffic crash and crime data of Taiwan and the United States, built up Data-Driven Approach to Crime and Traffic Safety (DDACTS) models and defined traffic crime hotspots. Other related studies include building up traffic crime prediction models, strategy evaluations of before-after traffic improvement, correlations of alcohol-related crashes with alcohol retailers and bar areas, spatial patterns of hit-and-run traffic crimes, and counseling demands of traffic crime victims. Most of our study results are presented at international conferences and published in SCI journals. Research Focus
Agent-Based Models
Spatial Statistics and GIS Applications Human Spatial Behaviors and Movements Spatial Data Mining and Modeling Spatial Statistics Publication
|
|
|