Iterative active feature extraction

Patent metadata

Field	Value
Publication number	US-9390377-B2
Application number	US-201313785132-A
Country	US
Kind code	B2
Filing date	Mar 5, 2013
Priority date	Dec 21, 2012
Publication date	Jul 12, 2016
Grant date	Jul 12, 2016

Abstract

Techniques for iterative feature extraction using domain knowledge are provided. In one aspect, a method for feature extraction is provided. The method includes the following steps. At least one query to predict at least one future value of a given value series based on a statistical model is received. At least two predictions of the future value are produced fulfilling at least the properties of 1) each being as probable as possible given the statistical model and 2) being mutually divert (in terms of numerical distance measure). A user is queried to select one of the predictions. The user may be queried for textual annotations for the predictions. The annotations may be used to identify additional covariates to create an extended set of covariates. The extended set of covariates may be used to improve the accuracy of the statistical model.

First claim

What is claimed is:

1. An apparatus for feature extraction, the apparatus comprising:
a memory; and
at least one processor device, coupled to the memory, operative to:
a) receive at least one query to predict at least one future value of a given value series;
b) generate a statistical model built on covariates to produce at least two predictions of the future value fulfilling at least the properties of 1) each being as statistically probable as possible given the statistical model wherein to be as statistically probable as possible an absolute distance of each of the predictions to a true value is less than a predetermined distance parameter with greater than a predetermined probability and 2) being mutually divert in terms of a numerical distance measure; and
c) query a user to select one of the predictions.

2. The apparatus of claim 1 , wherein the at least one processor device is further operative to:
d) query the user for textual annotations for the predictions;
e) use the annotations to identify additional covariates to create an extended set of covariates; and
f) use the extended set of covariates to improve the accuracy of the statistical model.

3. The apparatus of claim 1 , wherein the at least one processor device is further operative to:
present the predictions to the user; and
query the user to select which of the predictions the user believes is most probable.

4. The apparatus of claim 2 , wherein the textual annotations comprise tags.

5. The apparatus of claim 2 , wherein the at least one processor device is further operative to:
repeat the steps a-f in an iterative manner with the extended set of covariates.

6. The apparatus of claim 5 , wherein the at least one processor device is further operative to:
at each iteration, display the textual annotations from previous iterations with the predictions.

7. The apparatus of claim 1 , wherein the at least one processor device is further operative to:
obtain the statistical model.

8. The apparatus of claim 7 , wherein the statistical model is obtained from a statistician.

9. The apparatus of claim 1 , wherein the at least one processor device is further operative to:
recommend textual annotations from past interactions.

10. An article of manufacture for feature extraction, comprising a non-transitory machine-readable recordable medium containing one or more programs which when executed implement the steps of:
a) receiving at least one query to predict at least one future value of a given value series;
b) generating a statistical model built on covariates to produce at least two predictions of the future value fulfilling at least the properties of 1) each being as statistically probable as possible given the statistical model wherein to be as statistically probable as possible an absolute distance of each of the predictions to a true value is less than a predetermined distance parameter with greater than a predetermined probability and 2) being mutually divert in terms of a numerical distance measure; and
c) querying a user to select one of the predictions.

11. The article of manufacture of claim 10 , wherein the one or more programs which when executed further implement the steps of:
d) querying the user for textual annotations for the predictions;
e) using the annotations to identify additional covariates to create an extended set of covariates; and
f) using the extended set of covariates to improve the accuracy of the statistical model.

12. The article of manufacture of claim 10 , wherein the one or more programs which when executed further implement the steps of:
presenting the predictions to the user; and
querying the user to select which of the predictions the user believes is most probable.

13. The article of manufacture of claim 11 , wherein the textual annotations comprise tags.

14. The article of manufacture of claim 11 , wherein the one or more programs which when executed further implement the step of:
repeating the steps a-f in an iterative manner with the extended set of covariates.

15. The article of manufacture of claim 14 , wherein the one or more programs which when executed further implement the step of:
at each iteration, displaying the textual annotations from previous iterations with the predictions.

16. The article of manufacture of claim 10 , wherein the one or more programs which when executed further implement the step of:
obtaining the statistical model.

17. The article of manufacture of claim 16 , wherein the statistical model is obtained from a statistician.

18. The article of manufacture of claim 10 , wherein the one or more programs which when executed further implement the steps of:
recommending textual annotations from past interactions.

Assignees

• IBM

Inventors

• Lingenfelder Christoph
• Pompey Pascal
• Verscheure Olivier
• Wurst Michael

Classifications

• G06N5/043

  Distributed expert systems; Blackboards · CPC title

• G06N99/005 Primary

  Physics · mapped topic

• G06N5/02

  Knowledge representation; Symbolic representation · CPC title

• G06N5/025

  Extracting rules from data · CPC title

• G06N20/00 Primary

  Machine learning · CPC title