def rank(self):
X, Y = self.makeFeatures()
#pdb.set_trace()
X = np.asarray(X, dtype=np.float32)
Y = np.asarray(Y, dtype=np.int32)
X_train, X_test, y_train, y_test = train_test_split(X,
Y,
test_size=0.75,
random_state=0)
tuned_parameters = [{
'kernel': ['rbf'],
'gamma': [1e-3, 1e-2, 1e-1, 1, 10],
'C': [1e-2, 1e-1, 1, 10, 100, 1000]
}, {
'kernel': ['linear'],
'C': [1, 10, 100, 1000]
}]
clf = GridSearchCV(ranking.RankSVM(C=1), tuned_parameters, cv=5)
clf.fit(X_train, y_train)
print("Best parameters set found on development set:")
print(clf.best_params_)
print("Grid scores on development set:")
means = clf.cv_results_['mean_test_score']
stds = clf.cv_results_['std_test_score']
for mean, std, params in zip(means, stds, clf.cv_results_['params']):
print("%0.3f (+/-%0.03f) for %r" % (mean, std * 2, params))
be = clf.best_estimator_
def featureSelect(self):
X, y = self.makeFeatures()
X = np.asarray(X, dtype=np.float32)
y = np.asarray(y, dtype=np.int32)[:, 0]
svc = ranking.RankSVM(kernel='linear', C=10)
rfecv = RFECV(estimator=svc, step=1, cv=StratifiedKFold(2))
rfecv.fit(X, y)
print("Optimal number of features : %d" % rfecv.n_features_)
def rankSVM(self):
X, Y = self.makeFeatures()
X = np.asarray(X, dtype=np.float32)
Y = np.asarray(Y, dtype=np.int32)
y = Y
group = Y[:, 1]
gkf = GroupKFold(n_splits=10)
for train, test in gkf.split(X, y, groups=group):
svc = ranking.RankSVM(C=1e-1, kernel='rbf', gamma=10)
svc.fit(X[train], y[train])
print(svc.score(X[test], y[test]))
def svmRankingTrain(self):
X, Y = self.makeFeatures()
y = self.setLabel(Y)
X = np.asarray(X, dtype=np.float64)
Y = np.asarray(Y, dtype=np.int32)
group = Y[:, 1]
X_train, X_test, y_train, y_test = train_test_split(X,
y,
test_size=0.65,
random_state=0)
svc = ranking.RankSVM(C=1e1, kernel='rbf', gamma=1)
svc.fit(X_train, y_train)
y_out = svc.predict(X)
Y[:, 0] = y_out
ranks = self.relativeToAbsHalf(X, Y)
for i, j in enumerate(ranks):
self.pairs[i].append(j)