AIR QUALITY ANALYSIS BASED ON OUTLIER DETECTION ALGORITHMS

This study employs an outlier detection algorithm based on the distance metric of the Two-Step clustering algorithm to analyze potential untrustworthy data in pollutant concentration records across the region of Beijing-Tianjin-Hebei. By calculating anomaly indices through designated formulas and evaluating variable contribution rates, abnormal data points were identified for each monitoring area. Subsequent analysis of these anomalies provides substantial evidence supporting the existence of unreliable data within the dataset.