find_rate_variables_with_other_sampling_methods: Derive Rate Values
| find_rate_variables_with_other_sampling_methods | R Documentation |
Derive Rate Values
Description
find_rate_variables_with_other_sampling_methods returns a vector of variable importances using RATE.
We fit a GPC classifier to the data, draw samples from the latent posterior using one of Lapace Approximation, Elliptical Slice Samping,
or Expectation Propogation. After posterior inference, we fit RATE to derive association measure for each sub-level set of the design matrix.
Usage
find_rate_variables_with_other_sampling_methods(
gp_data,
bandwidth = 0.01,
type = "Laplace"
)
Arguments
gp_data |
(matrix) : The design matrix of (S/D) EC curves and the associated class labels |
bandwidth |
(float) : The bandwidth of the Gaussian Kernel used to fit the GPC. |
type |
(string) : The sampling method used. We currently support Laplace's method, Elliptical Slice Sampling, and Expectation Propogation. |
Value
rate_values (nx2 matrix) : The derived variable importance values and the row number denoting which sub-level set it corresponds to.
