Comparing event data sets

• US 9,372,664 B2
• Filed: 09/16/2014
• Issued: 06/21/2016
• Est. Priority Date: 04/20/2012
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Przestrzelski, Piotr, Thiebaut-George, Eric
• Primary Examiner(s): DO, AN H
• Application Number:
  US14/487,181
• Publication Number:
  US 20150006471A1
• Time in Patent Office: 644 Days
• Field of Search: 702/176, 702/178, 702/189, 707/603
• US Class Current: 1/1
• CPC Class Codes: G04F 13/00 Apparatus for measuring unknown time intervals by means not provided for in groups G04F5/00 - G04F10/00
• CPC Class Codes: G06F 17/10 Complex mathematical operations function generation by table look-up G06F1/03; evaluation of elementary functions by calculation G06F7/544
• CPC Class Codes: G06F 17/18 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 Class Codes: G06F 2218/08 Feature extraction
• CPC Class Codes: G06F 7/02 Comparing digital values G06F7/06, G06F7/22, G06F7/38 take precedence

Abstract

A method and apparatus for comparing event data sets. A method, system and computer program for comparing event data sets comprising: calculating one or more derivatives for each event in a first and second set of events; classifying each event by the calculated derivatives; and evaluating a similarity of the first and second sets of events based on the classifications of the respective events.

Claims

1. A method performed by a computer system for comparing event data sets comprising:

calculating, by the computer system, one or more derivatives for each event in a first and second set of events;

classifying, by the computer system, each event by respective calculated derivatives; and

correlating, by the computer system, the first and second sets of events based on the classifications of the respective events;

wherein each derivative is classified as a negative, zero or positive value and each event is classified according to a combination of negative, zero and positive values of the derivatives.

2. A method as claimed in claim 1, wherein a first derivative is approximated as a distance between two consecutive events divided by a difference of their time stamps.

3. A method as claimed in claim 2, wherein a second derivative is approximated as a distance between two consecutive first derivatives divided by the difference of the time stamps.

4. A method as claimed in claim 3, wherein a third derivative is approximated as a distance between two consecutive second derivatives divided by the difference of the time stamps.

5. A method as claimed in claim 2, wherein the difference in timestamps is a unit value.

6. A method as claimed in claim 2, further comprising a first derivative noise level as a fraction of a maximum derivative value whereby if the first derivative is between the absolute first derivative noise level then the first derivative is taken as neutral.

7. A method as claimed in claim 2, further comprising smoothing the metrics to reduce noise.

8. A method as claimed in claim 1, wherein correlation is determined if similarity of classification is above a threshold similarity.

9. A method as claimed in claim 1, wherein each of derivative classifications in a sequence is inverted and an anti-correlation is determined if the similarity of classification between the inverted sequence and a non-inverted sequence of events is above a threshold similarity.