Determining whether a point in a data stream is an outlier using hierarchical trees
AssigneeHewlett-Packard Development Company, L.P.
InventorsSong Wang, Chetan Kumar Gupta, Abhay Mehta
Life of this patent
Estimating the term…
Abstract
A technique that includes receiving a data stream that is indicative of a plurality of multi-dimensional points in a processor-based machine and for each dimension, organizing data indicative of values of the points in the dimension in an associated hierarchical tree. The technique includes using the processor-based machine to determine whether a given point of the plurality of points is an outlier based on a combination of the trees.
Claims20
A method comprising:
- receiving a data stream indicative of a plurality of multi-dimensional points in a processor-based machine;
- for each dimension, organizing data indicative of values of the points in the dimension in an associated hierarchical tree;
- and
- using the processor-based machine to determine whether a given point of the plurality of points is an outlier based on a combination of the trees.
A system comprising:
- a memory;
- and
- a processor-based outlier detector to:
- monitor a user specified time window of streaming data, the streaming data being indicative of a plurality of multi-dimensional points;
- for each of the dimensions, store data in the memory indicative of values of the points in the dimension in an associated hierarchical tree;
- and
- selectively combine the hierarchical trees based on a user specified set of dimensions to determine whether a given point of the plurality of points within the window is an outlier.
An article comprising a computer readable storage medium storing instructions that when executed by a computer cause the computer to:
- receive a data stream indicative of a plurality of multi-dimensional points;
- for each dimension, organize data indicative of values of the points in the dimension in an associated hierarchical tree;
- and
- determine whether a given point of the plurality of points is an outlier based on a combination of the trees.
Dependencies are inferred from the claim text.