Adjacent Mean Difference (AMD) method for dynamic segmentation in time series anomaly detection

摘要: Heuristic Ordered Time series Symbolic Aggregate approXimation (HOT SAX) is a well-known symbolic representation approach used to detect the abnormalities in time series. Since HOT SAX allows dimensionality reduction and searches with heuristic algorithm, prevail data analysis. However, because detects through sliding window of equal length, search results would change when setting different length optimal hard define. Therefore, this research, Adjacent Mean Difference (AMD) segmentation method was proposed segment dynamically without any parameter. Essentially, AMD partitions into multiple segments lengths based on transitions between points. After segmentation, FastDTW compare distances lengths. The experiments demonstrated that an easy efficient dynamically. And comparison shows can be better computational efficiency.