Using MASF
The Multivariate Adaptive Statistical Filtering (MASF) feature of Visualizer lets you display conventional Visualizer data in a new way. Instead of displaying measured values, MASF lets you display the difference between the measured values and your expectation of what the values should be. This makes it much easier to spot unusual values that you might want to investigate more thoroughly.
MASF is as valuable as the reference set you build. A reference set is a collection of measurement data obtained during multiple days of normal system operation. It's important to build a representative reference set.
MASF analyzes the values in the reference set and establishes expectations for the average and the degree of variability associated with each value. These expectations are referred to as a filtering policy.
Once MASF generates a filtering policy, you can apply that policy repeatedly to new sets of data by running a MASF event against the data. The result is that when you select a hierarchy with MASF graph from the Graphics menu, the selected graph looks like a standard hierarchy graph, but the values and colors in the graph reflect the results of MAS filtering rather than measured values.
You can define and apply multiple filtering policies as needed. For example, you might want to have filtering policies that reflect weekday expectations, ones that reflect weekends, or ones that reflect different shifts. You can update existing filtering policies by adding new data to existing reference sets.
This section describes the following information: