Pre-processed feature ranking for a support vector machine

• US 20050131847A1
• Filed: 11/07/2002
• Published: 06/16/2005
• Est. Priority Date: 05/01/1998
• Status: Active Grant

• Current Assignee: Curtis Anderson, Health Discovery Corporation, James Roberts, Joe Mckenzie, Jules B. Paderewski, Julian N. Stern, Memorial Health Systems Incorporated, Timothy P. O'Hayer
• Original Assignee: Curtis Anderson, Garry L. Carls, Glynn Bergeron, James Roberts, Joe Mckenzie, John E. Matthews, Jules B. Paderewski, Julian N. Stern, Memorial Health Systems Incorporated, Peter J. Farley, Simpson K. Russell, Timothy P. O'Hayer
• Inventors: Guyon, Isabelle, Perez-Cruz, Fernando, Elisseeff, Andre, Weston, Jason, Scholkopf, Bernhard
• Granted Patent: US 7,475,048 B2
• Time in Patent Office: Days
• US Class Current: 706/12
• CPC Class Codes: G06F 18/2113 by ranking or filtering the set of features, e.g. using a measure of variance or of feature cross-correlation
• CPC Class Codes: G06F 18/2411 based on the proximity to a decision surface, e.g. support vector machines
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 20/10 using kernel methods, e.g. support vector machines [SVM]

Abstract

Features are preprocessed (204) to minimize classification error in a Support Vector Machines (200) used to identify patterns in large databases. Pre-processing (204) is performed to constrain features used to train (210) the SVM learning machine. Live data (226) is collected and processed (232) with SVM.

Claims

1. A computer-implemented method for selecting a subset of features for processing in a learning machine, wherein the features correspond to a dataset to be analyzed for patterns, the method comprising:

ranking the features according to a distance between extremal points of two classes of interest; and

selecting the subset of features having the highest rank.

2. The method of claim 1, wherein the learning machine is a support vector machine.

3. The method of claim 1, further comprising pre-processing the dataset by normalizing the data.

4. The method of claim 3, wherein the step of pre-processing comprises:

arranging the data in a matrix having rows and columns;

first normalizing the columns of the matrix;

second normalizing the rows of the matrix; and

third normalizing the columns of the matrix.

5. The method of claim 1, wherein the dataset comprises gene expression data obtained from DNA micro-arrays.

6. The method of claim 5, wherein the gene expression data is obtained from tissue from patients with renal cancer.

7. A computer-implemented method for selecting a subset of features for processing in a learning machine, wherein the features correspond to a dataset to be analyzed for patterns, the method comprising:

determining a margin between extremal points of two classes of interest; and

ranking the subset of features according to the size of the margin, wherein the largest margin corresponds to the highest rank.

8. The method of claim 7, wherein the learning machine is a support vector machine.

9. The method of claim 7, further comprising pre-processing the dataset by normalizing the data.

10. The method of claim 9, wherein the step of pre-processing comprises:

arranging the data in a matrix having rows and columns;

first normalizing the columns of the matrix;

second normalizing the rows of the matrix; and

third normalizing the columns of the matrix.

11. The method of claim 7, wherein the dataset comprises gene expression data obtained from DNA micro-arrays.

12. The method of claim 11, wherein the gene expression data is obtained from tissue from patients with renal cancer.

13. A method for diagnosing renal cancer in a patient comprising:

entering gene expression data into a computer adapted for implementing a learning machine, the gene expressions data comprising gene expression levels in tissue obtained from the patient for a gene selected from the group consisting of small inducible cytokine A2 (monocyte chemotactic protein

     1) and ATP synthase, H+ transporting, mitochondrial F1 complex, alpha subunit, isoform 1, cardiac muscle.