Description Usage Arguments Examples
This is the confidence index weighted learning functions that takes patient's characteristic X, treatment A and outcome Y and estimate an optimal individual treament rule. Using cross-validation, the function will choose the hyper-parameter in rbf kernel and neigborhood definition automatically.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 |
H: |
n by P feature matrix |
A: |
treatment, takes value -1 and +1 with size n |
R2: |
outcome or residual vector with length n |
pi: |
propensity score with length n, when RCT, it shoud be set as 0.5 for each subject |
kernel: |
kernel used in SVM |
pentype: |
penalty in the residual estimation process, useful or RWL only |
XS: |
variable used in similarity calculation, if not provided, use all features |
sigmalst: |
candidate value for hyper-parameter sigma of rbf kernel |
cpar: |
candidate value for hyper-parameter C of rbf kernel |
theta_list: |
candidate value for hyper-parameter theta governing the definition of neighborhood |
frac.par: |
if "frac" similarity is used, then provide this fraction parameter, ranges from 0-1 |
method: |
either "cos" or "frac" for similarity definition |
m: |
number of fold for cross validation in parameter tunning, default is 10 |
e: |
least tolerated error, default is 1e-5 |
1 2 3 4 5 6 7 |
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.