State change detection and behavioral prediction of a nonlinear and non-stationary system

Patent metadata

Field	Value
Publication number	US-9978017-B1
Application number	US-201313970213-A
Country	US
Kind code	B1
Filing date	Aug 19, 2013
Priority date	Aug 19, 2013
Publication date	May 22, 2018
Grant date	May 22, 2018

Abstract

A method includes performing a comparison, at a computing device, of first probability distribution function (PDF) parameters of a first data set to second PDF parameters of a second data set. The first data set and the second data set at least partially overlap. The method also includes determining, at the computer device, based on the comparison whether a state change occurs between the first data set and the second data set. The method further includes, in response to determining that the state change occurs, generating an indication of the state change.

First claim

What is claimed is:

1. A method comprising:
performing a comparison, at a computing device, of first probability distribution function (PDF) parameters of a first data set to second PDF parameters of a second data set, wherein the first data set and second data set are subsets of a data output, and
wherein:
the first data set and the second data set partially overlap,
the first PDF parameters include first shape parameters indicating a first shape of a first PDF of the first data set,
the second PDF parameters include second shape parameters indicating a second shape of a second PDF of the second data set,
the first data set and the second data set are nonlinear and non-stationary and
the first data set and the second data set describes at least one of an amount, a frequency, or a location of particular events;

determining, at the computing device, bounding parameter values based on the first PDF parameters;
determining, at the computing device, that a state change occurs between the first data set and the second data set when values of the second PDF parameters are not within the bounding parameter values;
generating, at the computing device, an indication of the state change as third shape parameters to predict future values of a subsequent data set of the data output in response to determining that the state change occurs; and
displaying the indication at a display device.

2. The method of claim 1 , wherein determining that the state change occurs includes:
determining the first shape based on the first PDF parameters;
determining the second shape based on the second PDF parameters, wherein the first PDF parameters, the second PDF parameters, or a combination thereof are variable parameters;
determining a first vector based on the first shape, the first vector defined in a space having a number of dimensions equal to a number of the variable parameters;
determining a second vector in the space based on the second shape;
determining a difference in magnitude between the first vector and the second vector;
determining a difference in phase angle between the first vector and the second vector; and
generating a subsequent PDF based on the difference in magnitude and the difference in phase angle.

3. The method of claim 2 , wherein generating the subsequent PDF based on the difference in magnitude and the difference in phase angle includes:
extrapolating the difference in phase angle to determine a subsequent magnitude;
extrapolating the difference in phase angle to determine a subsequent phase angle; and
generating the subsequent PDF based on the subsequent magnitude and the subsequent phase angle.

4. The method of claim 3 , further comprising determining a threshold based on a change of total integrated probability, wherein the threshold is associated with the state change.

5. The method of claim 4 , wherein a particular state change is detected when an integrated value of the second PDF is greater than an integrated value of the first PDF by an amount equal to or greater than the threshold.

6. The method of claim 2 , further comprising determining a projected range of data points of the subsequent data set when the state change occurs.

7. The method of claim 6 , wherein determining the projected range of data points of the subsequent data set based on the state change includes integrating the subsequent PDF within particular probability limits.

8. The method of claim 6 , wherein the projected range of data points of the subsequent data set is determined without a noise reduction operation.

9. The method of claim 6 , wherein the subsequent PDF is determined using a continuous probability distribution function.

10. The method of claim 9 , wherein the continuous probability distribution function is a Weibull PDF, and wherein the first PDF parameters and the second PDF parameters correspond to parameters of the Weibull PDF.

11. The method of claim 1 , further comprising:
determining subsequent PDF parameters of the subsequent data set when the state change occurs; and
determining a projected range of data points of the subsequent data set based on the subsequent PDF parameters.

12. The method of claim 11 , wherein the subsequent PDF parameters are determined using a Bayesian decision process, and wherein a number of the projected range of data points is equal to a difference between a number of data points in the first data set and a number of overlapped data points.

13. The method of claim 1 , wherein the subsequent data set is subsequent in time or space to the first data set and the second data set, and wherein the second data set overlaps the first data set by at least half of data points in the first data set.

14. The method of claim 1 , wherein the first data set and the second data set describe at least one of network traffic, geographic locations of political events, temperature induced mode changes in electronic equipment, crime rates within bounded areas, or a combination thereof, and wherein the state change is determined without using a priori information related to the first data set.

15. An apparatus comprising:
a processor; and
a memory coupled to the processor, wherein the memory stores instructions executable by the processor to perform operations comprising:
performing a comparison, by the processor, of first probability distribution function (PDF) parameters of a first data set to second PDF parameters of a second data set, wherein the first data set and second data set are subsets of a data output, and wherein the first PDF parameters include first shape parameters indicating a first shape of a first PDF of the first data set and the second PDF parameters include second shape parameters indicating a second shape of a second PDF of the second data set, the first data set and the second data set at least partially overlap, the first data set and the second data set are nonlinear and non-stationary, and the first data set and the second data set describes at least one of an amount, a frequency, or a location of particular events;
determining bounding parameter values based on the first PDF parameters;
determining that a state change occurs between the first data set and the second data set when values of the second PDF parameters are not within the bounding parameter values;
generating an indication of the state change to predict future values of a subsequent data set of the data output as third shape parameters in response to determining that the state change occurs; and
displaying the indication at a display device.

16. The apparatus of claim 15 , wherein the first data set and the second data set describe at least one of network traffic, geographic locations of political events, temperature induced mode changes in electronic equipment, crime rates within bounded areas, or a combination thereof, and wherein determining that the state change occurs includes:
determining the first shape based on the first PDF parameters;
determining second shape based on the second PDF parameters, wherein at least one of the first PDF parameters and the second PDF parameters are variable parameters;
determining a first vector based on the first shape, the first vector defined in a space having a number of dimensions equal to a number of the variable parameters;
determining a second vector in the space based on the second shape;
determining a difference in magnitude between the first vector and the second vector;
determining a difference in phase angle between the first vector and the second vector; and
generating a subsequent PDF based on the differen

Assignees

• Boeing Co

Inventors

• Simms Donald S

Classifications

• G06N7/01 Primary

  Probabilistic graphical models, e.g. probabilistic networks · CPC title

• G06F17/18 Primary

  for evaluating statistical data {, e.g. average values, frequency distributions, probability functions, regression analysis (forecasting specially adapted for a specific administrative, business or logistic context G06Q10/04)} · CPC title

• G06N5/02 Primary

  Knowledge representation; Symbolic representation · CPC title