ENERGY MINIMIZATION FOR CLASSIFICATION, PATTERN RECOGNITION, SENSOR FUSION, DATA COMPRESSION, NETWORK RECONSTRUCTION AND SIGNAL PROCESSING

• US 20100232718A1
• Filed: 04/09/2010
• Published: 09/16/2010
• Est. Priority Date: 12/29/1997
• Status: Active Grant

• Current Assignee: Abel Wolman, Jeff B. Glickman
• Original Assignee: Abel Wolman, Jeff B. Glickman
• Inventors: Wolman, Abel, Glickman, Jeff B.
• Granted Patent: US 7,912,290 B2
• Time in Patent Office: Days
• US Class Current: 382/225
• CPC Class Codes: G06F 18/2137 based on criteria of topology preservation, e.g. multidimensional scaling or self-organising maps
• CPC Class Codes: Y10S 707/99932 Access augmentation or optimizing
• CPC Class Codes: Y10S 707/99942 Manipulating data structure, e.g. compression, compaction, compilation

Abstract

A data analyzer/classifier comprises using a preprocessing step, and energy minimization step, and a postprocessing step to analyze/classify data.

Claims

1. A method for determining dimensionality of a network, the dimensionality corresponding to a number of degrees of freedom in the network, the method comprisings:

using a computer system, sampling data from one or more nodes of the network;

using a computer system, mapping the data into one or more matrices;

using a computer system, applying individual differences multidimensional scaling to the one or more matrices to produce a common space output; and

using a computer system, processing the common space output to determine the dimensionality of the network.

2. The method for determining dimensionality of a network of claim 1 wherein mapping the data into one or more matrices comprises a processing step selected from the group consisting of forming one or more symmetric matrices, forming one or more hollow symmetric matrices, entry-wise substituting the received data to populate one or more symmetric matrices, forming one or more proximity matrices, forming one or more distance matrices, and forming one or more Euclidean distance matrices.

3. The method for determining dimensionality of a network of claim 1 wherein applying individual differences multidimensional scaling comprises applying nonmetric individual differences multidimensional scaling.

4. A method for reconstructing a network, the method comprising:

using a computer system, sampling data from one or more nodes of the network;

using a computer system, mapping the data into one or more matrices;

using a computer system, applying individual differences multidimensional scaling to the one or more matrices to produce a source space output;

from the source space output, using a computer system, determining the dimensionality of the network;

using a computer system, using free nodes to recreate and reconstruct individual nodes through the use of matrices containing missing values; and

using a computer system, establishing node connectivity through the use of lowest-energy connections constrained by dimensionality.

5. The method for reconstructing a network of claim 4 wherein mapping the data into one or more matrices comprises a processing step selected from the group consisting of forming one or more symmetric matrices, forming one or more hollow symmetric matrices, entry-wise substituting the received data to populate one or more symmetric matrices, forming one or more proximity matrices, forming one or more distance matrices, and forming one or more Euclidean distance matrices.

6. The method for reconstructing a network of claim 4 wherein applying applying individual differences multidimensional scaling comprises applying nonmetric individual differences multidimensional scaling.

7. A method for determining dimensionality of a dynamical system from partial data, the dimensionality corresponding to a number of degrees of freedom in the dynamical system, the method comprising:

using a computer system, sampling data from the dynamical system;

using a computer system, mapping the data into one or more matrices;

using a computer system, applying individual differences multidimensional scaling to the one or more matrices to produce a common space output;

using a computer system, processing the common space output to determine dimensionality of the dynamical system.

8. The method determining dimensionality of a dynamical system from partial data of claim 7 wherein mapping the data into one or more matrices comprises a processing step selected from the group consisting of forming one or more symmetric matrices, forming one or more hollow symmetric matrices, entry-wise substituting the received data to populate one or more symmetric matrices, forming one or more proximity matrices, forming one or more distance matrices, and forming one or more Euclidean distance matrices.

9. The method for determining dimensionality of a dynamical system from partial data of claim 7 wherein applying individual differences multidimensional scaling comprises applying nonmetric individual differences multidimensional scaling.