Support vector regression for censored data

• US 20050108753A1
• Filed: 11/17/2004
• Published: 05/19/2005
• Est. Priority Date: 11/18/2003
• Status: Active Grant

• Current Assignee: David Sans
• Original Assignee: Aureon Biosciences, Inc.
• Inventors: Verbel, David A., Saidi, Olivier
• Granted Patent: US 7,505,948 B2
• Time in Patent Office: Days
• US Class Current: 725/46
• 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 18/2178 based on feedback of a supervisor
• CPC Class Codes: G06F 18/2411 based on the proximity to a decision surface, e.g. support vector machines
• CPC Class Codes: G06V 10/764 using classification, e.g. of video objects
• CPC Class Codes: G06V 10/7784 based on feedback from supervisors
• CPC Class Codes: G16H 10/20 for electronic clinical trials or questionnaires

Abstract

A method of producing a model for use in predicting time to an event includes obtaining multi-dimensional, non-linear vectors of information indicative of status of multiple test subjects, at least one of the vectors being right-censored, lacking an indication of a time of occurrence of the event with respect to the corresponding test subject, and performing regression using the vectors of information to produce a kernel-based model to provide an output value related to a prediction of time to the event based upon at least some of the information contained in the vectors of information, where for each vector comprising right-censored data, a censored-data penalty function is used to affect the regression, the censored-data penalty function being different than a non-censored-data penalty function used for each vector comprising non-censored data.

Claims

1. A method of producing a model for use in predicting time to an event, the method comprising:

obtaining multi-dimensional, non-linear vectors of information indicative of status of multiple test subjects, at least one of the vectors being right-censored, lacking an indication of a time of occurrence of the event with respect to the corresponding test subject; and

performing regression using the vectors of information to produce a kernel-based model to provide an output value related to a prediction of time to the event based upon at least some of the information contained in the vectors of information;

wherein for each vector comprising right-censored data, a censored-data penalty function is used to affect the regression, the censored-data penalty function being different than a non-censored-data penalty function used for each vector comprising non-censored data.

2. The method of claim 1 wherein the regression comprises support vector machine regression.

3. The method of claim 1 wherein the censored-data penalty function has a larger positive slack variable than the non-censored data penalty function does.

4. The method of claim 1 wherein performing the regression includes using penalty functions that include linear functions of a difference between a predicted value of the model and a target value for the predicted value, and wherein a first slope of the linear function for positive differences between the predicted and target values for the censored-data penalty function is lower than a second slope of the linear function for positive differences between the predicted and target values for the non-censored-data penalty function.

5. The method of claim 4 wherein the first slope is substantially equal to a third slope of the linear function for negative differences between the predicted and target values for the censored-data penalty function and a fourth slope of the linear function for negative differences between the predicted and target values for the non-censored-data penalty function, and wherein positive and negative slack variables of the non-censored-data penalty function and a negative slack variable of the censored-data penalty function are substantially equal.

6. The method of claim 1 wherein the data of the vectors are associated with categories based on at least one characteristic of the data that relate to the data'"'"'s ability to help the model provide the output value such that the output value helps predict time to the event, the method further comprising performing the regression using the data from the vectors in sequence from the category with data most likely, to the category with data least likely, to help the model provide the output value such that the output value helps predict time to the event.

7. The method of claim 6 wherein the at least one characteristic is at least one of reliability and predictive power.

8. The method of claim 6 wherein the regression is performed in a greedy-forward manner in accordance with the features of the data to select features to be used in the model.

9. The method of claim 8 further comprising performing a greedy backward procedure to the features of the vectors, after performing the regression, to further select features to be used in the model.

10. The method of claim 8 wherein the regression is performed in the greedy-forward manner with respect to only a portion of the features of the vectors.

11. The method of claim 10 wherein the vectors include categories of data of clinical/histopathological data, biomarker data, and bio-image data, and wherein the regression is performed in the greedy-forward manner with respect to only the biomarker data and the bio-image data of the vectors.

12. The method of claim 1 wherein the vectors of information are indicative of status of test subjects that are at least one of living, previously-living, and inanimate.

13. A computer program product producing a model for use in predicting time to an event, the computer program product residing on a computer readable medium, the computer program product comprising computer-readable, computer-executable instructions for causing a computer to:

obtain multi-dimensional, non-linear vectors of information indicative of status of multiple test subjects, at least one of the vectors being right-censored, lacking an indication of a time of occurrence of the event with respect to the corresponding test subject; and

perform regression using the vectors of information to produce a kernel-based model to provide an output value related to a prediction of time to the event based upon at least some of the information contained in the vectors of information;

wherein for each vector comprising right-censored data, a censored-data penalty function is used to affect the regression, the censored-data penalty function being different than a non-censored-data penalty function used for each vector comprising non-censored data.

14. The computer program product of claim 13 wherein the regression comprises support vector machine regression.

15. The computer program product of claim 13 wherein the censored-data penalty function has a larger positive slack variable than the non-censored data penalty function does.

16. The computer program product of claim 13 wherein the instructions for causing the computer to perform the regression include instruction for causing the computer to use penalty functions that include linear functions of a difference between a predicted value of the model and a target value for the predicted value, and wherein a first slope of the linear function for positive differences between the predicted and target values for the censored-data penalty function is lower than a second slope of the linear function for positive differences between the predicted and target values for the non-censored-data penalty function.

17. The computer program product of claim 16 wherein the first slope is substantially equal to a third slope of the linear function for negative differences between the predicted and target values for the censored-data penalty function and a fourth slope of the linear function for negative differences between the predicted and target values for the non-censored-data penalty function, and wherein positive and negative slack variables of the non-censored-data penalty function and a negative slack variable of the censored-data penalty function are substantially equal.

18. The computer program product of claim 13 wherein the instructions for causing the computer to perform regression cause the regression to be performed using the data from the vectors in sequence from a category with data most likely, to a category with data least likely, to help the model provide the output value such that the output value helps predict time to the event.

19. The computer program product of claim 18 wherein the instructions for causing the computer to perform regression cause the regression to be performed in a greedy-forward manner in accordance with features of the data to select features to be used in the model.

20. The computer program product of claim 19 further comprising instructions for causing the computer to perform a greedy backward procedure to the features of the model, after performing the regression, to further select features to be used in the model.

21. The computer program product of claim 19 wherein the instructions for causing the computer to perform regression in the greedy-forward manner cause the computer to perform the greedy-forward feature selection with respect to only a portion of the features of the vectors.

22. The computer program product of claim 21 wherein the vectors include categories of data of clinical/histopathological data, biomarker data, and bio-image data, and wherein the instructions for causing the computer to perform regression in the greedy-forward manner cause the computer to perform the greedy-forward feature selection with respect to only the biomarker data and the bio-image data of the vectors.

23. A method of producing a model for use in predicting time to an event, the method comprising:

obtaining multi-dimensional, non-linear vectors of information indicative of status of multiple test subjects; and

performing regression using the vectors of information to produce a kernel-based model to provide an output value related to a prediction of time to the event based upon at least some of the information contained in the vectors of information;

wherein the data of the vectors are associated with categories based on at least one characteristic of the data that relate to the data'"'"'s ability to help the model provide the output value such that the output value helps predict time to the event; and

wherein the regression is performed using the data from the vectors in sequence from the category with data most likely, to the category with data least likely, to help the model provide the output value such that the output value helps predict time to the event.

24. The method of claim 23 wherein the regression is performed in a greedy-forward manner in accordance with features of the data to select features to be used in the model.

25. The method of claim 24 further comprising performing a greedy backward procedure to the features of the vectors, after performing the regression, to further select features to be used in the model.

26. The method of claim 24 wherein the regression is performed in the greedy-forward manner with respect to only a portion of the features of the vectors.

27. The method of claim 26 wherein the vectors include categories of data of clinical/histopathological data, biomarker data, and bio-image data, and wherein the regression is performed in a non-greedy-forward manner with the clinical/histopathological data and in the greedy-forward manner with respect to only the biomarker data and the bio-image data of the vectors, in that order.

28. The method of claim 23 wherein at least one of the vectors is right-censored, lacking an indication of a time of occurrence of the event with respect to the corresponding test subject.

29. A computer program product for producing a model for use in predicting time to an event, the computer program product residing on a computer readable medium and comprising computer-readable, computer-executable instructions for causing a computer to:

obtain multi-dimensional, non-linear vectors of information indicative of status of multiple test subjects, at least one of the vectors being right-censored, lacking an indication of a time of occurrence of the event with respect to the corresponding test subject; and

perform regression using the vectors of information to produce a kernel-based model to provide an output value related to a prediction of time to the event based upon at least some of the information contained in the vectors of information;

wherein the data of the vectors are associated with categories based on at least one characteristic of the data that relate to the data'"'"'s ability to help the model provide the output value such that the output value helps predict time to the event; and

wherein the regression is performed using the data from the vectors in sequence from the category with data most likely, to the category with data least likely, to help the model provide the output value such that the output value helps predict time to the event.

30. The computer program product of claim 29 wherein the regression is performed in a greedy-forward manner in accordance with features of the data to select features to be used in the model.

31. The computer program product of claim 30 further comprising instructions for causing the computer to perform a greedy backward procedure to the features of the vectors, after performing the regression, to further select features to be used in the model.

32. The computer program product of claim 30 wherein the regression is performed in the greedy-forward manner with respect to only a portion of the features of the vectors.

33. The computer program product of claim 32 wherein the vectors include categories of data of clinical/histopathological data, biomarker data, and bio-image data, and wherein the regression is performed in a non-greedy-forward manner with the clinical/histopathological data and in the greedy-forward manner with respect to only the biomarker data and the bio-image data of the vectors, in that order.

34. A method of determining a predictive diagnosis for a patient, the method comprising:

receiving at least one of clinical and histopathological data associated with the patient;

receiving biomarker data associated with the patient;

receiving bio-image data associated with the patient; and

applying at least a portion of the at least one of clinical and histopathological data, at least a portion of the biomarker data, and at least a portion of the bio-image data to a kernel-based mathematical model to calculate a value indicative of a diagnosis for the patient.

35. The method of claim 34 wherein the at least a portion of the biomarker data comprises data for less than all biomarker features of the patient.

36. The method of claim 35 wherein the at least a portion of the biomarker data comprises data for less than about ten percent of all biomarker features of the patient.

37. The method of claim 36 wherein the at least a portion of the biomarker data comprises data for less than about five percent of all biomarker features of the patient.

38. The method of claim 34 wherein the at least a portion of the biomarker data comprises data for less than all bio-image features of the patient.

39. The method of claim 38 wherein the at least a portion of the biomarker data comprises data for less than about one percent of all bio-image features of the patient.

40. The method of claim 39 wherein the at least a portion of the biomarker data comprises data for less than about 0.2 percent of all bio-image features of the patient.

41. The method of claim 34 wherein the value is indicative of at least one of a time to recurrence of a health-related condition and a probability of recurrence of the health-related condition.

42. An apparatus for determining time-to-event predictive information, the apparatus comprising:

an input configured to obtain multi-dimensional, non-linear first data associated with a possible future event; and

a processing device configured to use the first data in a kernel-based mathematical model, derived at least partially from a regression analysis of multi-dimensional, non-linear, right-censored second data that determines parameters of the model that affect calculations of the model, to calculate the predictive information indicative of at least one of a predicted time to the possible future event and a probability of the possible future event.

43. The apparatus of claim 42 wherein the input and the processing device comprise portions of a computer program product residing on a computer readable medium, the computer program product comprising computer-readable, computer-executable instructions for causing a computer to obtain the first data and to use the first data in the mathematical model to calculate the predictive information.

44. The apparatus of claim 42 wherein the first data comprises at least one of clinical and histopathological data, biomarker data, and bio-image data associated with a patient, and wherein the processing device is configured to use at least a portion of the at least one of clinical and histopathological data, at least a portion of the biomarker data, and at least a portion of the bio-image data to a kernel-based mathematical model to calculate the predictive information for the patient.

45. The apparatus of claim 44 wherein the at least a portion of the biomarker data comprises data for less than all biomarker features of the patient.

46. The apparatus of claim 45 wherein the at least a portion of the biomarker data comprises data for less than about five percent of all biomarker features of the patient.

47. The apparatus of claim 44 wherein the at least a portion of the biomarker data comprises data for less than all bio-image features of the patient.

48. The apparatus of claim 47 wherein the at least a portion of the biomarker data comprises data for less than about 0.2 percent of all bio-image features of the patient.

49. A computer program product for determining a predictive diagnosis for a patient, the computer program product residing on a computer readable medium and comprising computer-readable, computer-executable instructions for causing a computer to:

receive at least one of clinical and histopathological data associated with the patient;

receive biomarker data associated with the patient;

receive bio-image data associated with the patient; and

apply at least a portion of the at least one of clinical and histopathological data, at least a portion of the biomarker data, and at least a portion of the bio-image data to a kernel-based mathematical model to calculate a value indicative of a diagnosis for the patient.

50. The computer program product of claim 49 wherein the at least a portion of the biomarker data comprises data for less than all biomarker features of the patient.

51. The computer program product of claim 50 wherein the at least a portion of the biomarker data comprises data for less than about ten percent of all biomarker features of the patient.

52. The computer program product of claim 51 wherein the at least a portion of the biomarker data comprises data for less than about five percent of all biomarker features of the patient.

53. The computer program product of claim 49 wherein the at least a portion of the biomarker data comprises data for less than all bio-image features of the patient.

54. The computer program product of claim 53 wherein the at least a portion of the biomarker data comprises data for less than about one percent of all bio-image features of the patient.

55. The computer program product of claim 54 wherein the at least a portion of the biomarker data comprises data for less than about 0.2 percent of all bio-image features of the patient.

56. The computer program product of claim 49 wherein the value is indicative of at least one of a time to recurrence of a health-related condition and a probability of recurrence of the health-related condition.