predicting_time_targets = 0.0
# grid search
for index, i in enumerate(nu_list):
for jndex, j in enumerate(gamma_list):
print("nu=%r, gamma=%r"%(i,j))
# model fitting
start = timeit.default_timer()
clf = ocsvm.fit(new_data, i, j)
stop = timeit.default_timer()
training_time_sum += stop - start
# predicting
start = timeit.default_timer()
y_outliers = ocsvm.predict(clf, pseudo_outliers)
stop = timeit.default_timer()
predicting_time_outliers += stop - start
start = timeit.default_timer()
y_targets = ocsvm.predict(clf, pseudo_targets)
stop = timeit.default_timer()
predicting_time_targets += stop - start
# calculate the error
err_outliers = float(y_outliers[y_outliers[:]==1].shape[0]) / y_outliers.shape[0]
err_targets = float(y_targets[y_targets[:]!=1].shape[0]) / y_targets.shape[0]
err = 0.5*err_outliers + 0.5*err_targets
# save results
error_array[index][jndex] = err
full_err_array[index][jndex] = [err_outliers, err_targets]
