Time Series Based Data Explorer and Stream Analysis for Anomaly Prediction

Figure 2

An example of phased anomaly prediction with feature subsequence set as a precursor. In this monitoring system based on a three-dimensional time series, anomalous antecedents are divided into three phases, which are characterized by feature subsequence sets {, },{}, and {, , }.