Clustering as an EDA method: the case of pedestrian directional flow behavior
Given the data of pedestrian trajectories in NTXY format, three clustering methods of K Means, Expectation Maximization (EM) and Affinity Propagation were utilized as Exploratory Data Analysis to find the pattern of pedestrian directional flow behavior. The analysis begins without a prior notion reg...
- Autores:
-
Teknomo, Kardi
Estuar, Ma. Regina
- Tipo de recurso:
- Fecha de publicación:
- 2010
- Institución:
- Universidad de San Buenaventura
- Repositorio:
- Repositorio USB
- Idioma:
- spa
- OAI Identifier:
- oai:bibliotecadigital.usb.edu.co:10819/6449
- Acceso en línea:
- http://hdl.handle.net/10819/6449
- Palabra clave:
- Gaussian Mixture
Directional flow pattern
Pedestrian behavior
Trajectory analysis
Mezcla Gaussiana
Patrón de flujo direccional
Comportamiento peatonal
Análisis de trayectoria
Análisis de datos
- Rights
- License
- Atribución-NoComercial-SinDerivadas 2.5 Colombia
| Summary: | Given the data of pedestrian trajectories in NTXY format, three clustering methods of K Means, Expectation Maximization (EM) and Affinity Propagation were utilized as Exploratory Data Analysis to find the pattern of pedestrian directional flow behavior. The analysis begins without a prior notion regarding the structure of the pattern and it consequentially infers the structure of directional flow pattern. Significant similarities in patterns for both individual and instantaneous walking angles based on EDA method are reported and explained in case studies. |
|---|