| Title |
Comparison Of Outlier Detection Approaches In A Smart Cities Sensor Data Context |
| ID_Doc |
15167 |
| Authors |
Zafeirelli S.; Kavroudakis D. |
| Year |
2024 |
| Published |
International Journal on Smart Sensing and Intelligent Systems, 17, 1 |
| DOI |
http://dx.doi.org/10.2478/ijssis-2024-0004 |
| Abstract |
This study examines outlier detection in time-series sensor data from PurpleAir low-cost sensors in Athens, Greece. Focusing on key environmental parameters such as temperature, humidity, and particulate matter (PM) levels, the study utilizes the Interquartile Range (IQR) and Generalized Extreme Studentized Deviate (GESD) methods on hourly and daily basis. GESD detected more outliers than IQR, most of them in PM, while temperature and humidity data had fewer outliers; applying filters before outlier detection and adjusting alpha values based on time scales were crucial, and outliers significantly affected spatial interpolation, emphasizing the need for spatial statistics in smart city air quality management. © 2024 Sciendo. All rights reserved. |
| Author Keywords |
environmental data; IoT; Outlier detection; sensors; smart cities |
