A Log-Based Anomaly Detection Method with Efficient Neighbor Searching and Automatic K Neighbor Selection

Figure 1

The influence of different k values on classification results.