Gecko Modeling

Modeling is the process of condensing the raw data into a terse concise model.  SensorMINER utilizes several algorithms to perform this including rule induction, path modeling, and box modeling. Each method has its own advantages and disadvantages producing specific output. Regardless of the modeling method used, the common goal is to build the necessary files required by the anomaly detection engine.

Gecko is an algorithm used for clustering time series data and automatically determines a reasonable number of clusters (states). The algorithm consists of a top-down partitioning phase to find initial sub-clusters and a bottom-up phase which merges them back together. The appropriate number of clusters is automatically determined by what we call the “L” method. To characterize the states in logical rules, we use the rule induction learning algorithm. To track normal behavior and detect anomalies, we construct a finite state automaton (FSA) with the identified states. In summary, Gecko is used for clustering time series data, we then integrate rule induction and state transition logic to generate a complete anomaly detection system.

sensorMiner Gecko Clustering

Click here for more information on Gecko Modeling: 

Learning States for Detecting Anomalies in Time Series

Determining the Number of Clusters/Segments in Hierarchical Clustering/Segmentation Algorithms

Learning States and Rules for Time Series Anomaly Detection