Method for selecting features used in continuous-valued regression analysis

1. A method for selecting features used in continuous-valued regression analysis, comprising the steps of:

providing a training data set as input, wherein the training data set includes features and corresponding target values, wherein the target values are continuous, and there is one target value for each feature;

thresholding and discretizing each target value with respect to a threshold value to produce a discretized target value; and

applying categorical feature selection to the features using the discretized target values to produces selected features, wherein the steps are performed in a processor.

2. The method of claim 1 , further comprising:

performing continuous-valued regression analysis using the selected features.

3. The method of claim 1 , wherein the thresholding and discretizing is with respect to one or more threshold values into which the target values are partitioned.

4. The method of claim 3 , further comprising:

selecting a single threshold value to be a median of the target values in a training data set to results in balanced classes.

5. The method of claim 2 , wherein the one or more threshold values are based on application-specific knowledge.

6. The method of claim 1 , wherein the regression analysis is nonlinear.

7. The method of claim 1 , wherein the regression analysis is a heteroscedastic support vector regression with least absolute deviation and l 1 regularization.