Ranking Support Vector Machine.
A trained ranking SVM model will predict rank-preserving values for each
sample.
The constraint violation cost will be computed according to
\[C_i = 10^{C_{base}} (N_{training} - i)^{C_{pow}},
\text{ for} i = 1, \ldots, N_{training}\]
so that top-ranked samples have higher violation costs.
| Parameters: |
- n_iter : int, optional (default: 50000 * sqrt(n_features))
Number of training iterations
- epsilon : float, optional (default: 1.0)
Tolerance ranking value error
- c_base : float, optional (default: 6)
Base for constraint violation cost
- c_pow : float, optional (default: 2)
Exponent for constraint violation cost
- c_sigma : float, optional (default: 1)
The sigma of the RBF kernel will be set to c_sigma times the average
distance of training samples
- random_state : optional, int
Seed for the random number generator
|
|---|
-
__init__(n_iter=-1, epsilon=1.0, c_base=6.0, c_pow=2.0, c_sigma=1.0, random_state=None)[source]
-
fit(X)[source]
Fit ranking SVM.
| Parameters: |
- X : array, shape (n_samples, n_features)
Training data, sorted, highest rank first
|
|---|
-
predict(X)[source]
Predict ranking values for new data.
| Parameters: |
- X : array, shape (n_test, n_features)
Test data
|
|---|
| Returns: |
- y : array, shape (n_test,)
Ranking values
|
|---|
