×

PREDICTING CHANGE-POINTS IN CONTINUOUS TIME-SERIES DATA

  • US 20140136454A1
  • Filed: 10/21/2013
  • Published: 05/15/2014
  • Est. Priority Date: 11/09/2012
  • Status: Active Grant
First Claim
Patent Images

1. A method for predicting a change-point in continuous time-series data by computer, the method comprising:

  • accepting an input of continuous time-series data in past w periods; and

    determining whether or not a change-point in time-series appears in future k periods by;

    determining, by a plurality of preliminary classifiers on the basis of the continuous time-series data in the past w periods and their respective preliminary classification rules, whether or not the change-point appears in future (k+i) periods, where i=0, 1, . . . , m; and

    determining, by a secondary classifier on the basis of determinations made by the plurality of preliminary classifiers and a secondary classification rule, whether or not the change-point appears in future c periods.

View all claims
  • 1 Assignment
Timeline View
Assignment View
    ×
    ×