Back to results
US 8,898,093 B1Granted
Systems and methods for analyzing data using deep belief networks (DBN) and identifying a pattern in a graph
AssigneeThe Boeing Company
InventorJohn Helmsen
Life of this patent
Estimating the term…
Abstract
A method for analyzing data is provided. The method includes generating, using a processing device, a graph from raw data, the graph including a plurality of nodes and edges, deriving, using the processing device, at least one label for each node using a deep belief network, and identifying, using the processing device, a predetermined pattern in the graph based at least in part on the labeled nodes.
Claims20
A method for analyzing data, said method comprising:
- generating, using an analyst assistant, a graph from raw data, the graph including a plurality of nodes and edges, wherein each node represents a piece of data and each edge represents a relationship between connected nodes;
- deriving, using the analyst assistant, at least one label for each node using a deep belief network;
- and
- identifying, using the analyst assistant, a predetermined pattern in the graph based at least in part on the labeled nodes.
1 more element
A data analysis system comprising:
- a memory device configured to store raw data;
- and
- an analyst assistant in communication with said memory device and configured to:
- generate a graph from the raw data, the graph including a plurality of nodes and edges, wherein each node represents a piece of data and each edge represents a relationship between connected nodes;
- derive at least one label for each node using a deep belief network;
- and
- identify a predetermined pattern in the graph based at least in part on the labeled nodes.
4 more elements
An analyst assistant embodied in a non-transitory computer readable medium, said analyst assistant configured to:
- generate a graph from raw data, the graph including a plurality of nodes and edges, wherein each node represents a piece of data and each edge represents a relationship between connected nodes;
- derive at least one label for each node using a deep belief network;
- and
- identify a predetermined pattern in the graph based at least in part on the labeled nodes.
1 more element
Dependencies are inferred from the claim text.