Anomaly detection method based on IoT and apparatus thereof

An anomaly detection apparatus based on IoT, the anomaly detection apparatus comprising:

• a principal component analysis unit configured to search for one principal component axis maximizing variance of data by analyzing a normal data set collected in time series from a plurality of IoT devices by using a principal component analysis technique;
• a center point setting unit configured to set a center point of the principal component axis based on a result of projecting a plurality of normal data which is each element in the normal data set onto the principal component axis;
• an input unit configured to receive a currently measured measurement data set from the plurality of IoT devices;
• a linear transformation unit configured to acquire a linear transformation data set having a plurality of projection points as elements by projecting a plurality of measurement data which is each element in the measurement data set onto the principal component axis;
• a computation unit configured to calculate a Mahalanobis distance between the projection point which is each element in the linear transformation data set, and the set central point;
• and
• a detection unit configured to detect whether or not data of the IoT devices is abnormal by comparing the Mahalanobis distance calculated for each element with a threshold.