An Efficient Outlier Detection Approach for Streaming Sensor Data Based on Neighbor Difference and Clustering

Table 2

The detection accuracy of the ODNDC approach.

Types	Variables
	Temperature (%)	Humidity (%)	Atmospheric pressure
Point outliers	83.3	93.8	—
Collective outliers	93.7	94.9	94.8%
Jump outliers	100	93.3	100%
Contextual outliers	96.4	93.1	93.1%