k = pyGPs.cov.RBF()
model.setPrior(kernel=k)
# split training and test data
x_train = x[index_train, :]
y_train = y[index_train, :]
x_test = x[index_test, :]
y_test = y[index_test, :]
# gp
model.optimize(x_train, y_train)
model.predict(x_test)
# evaluation
predictive_class_rbf = np.sign(model.ym)
ACC_rbf = valid.ACC(predictive_class_rbf, y_test)
## DIFFUSION Kernel
# compute kernel matrix and initalize GP with precomputed kernel
model = pyGPs.GPC()
M1, M2 = graphUtil.formKernelMatrix(Matrix, index_train, index_test)
k = pyGPs.cov.Pre(M1, M2)
model.setPrior(kernel=k)
# if you only use precomputed kernel matrix, there is no training data needed,
# but you still need to specify x_train (due to general structure of pyGPs)
# e.g. you can use the following:
n = len(index_train)
x_train = np.zeros((n, 1))
# gp
# normalize kernel matrix (not useful for MUTAG)
# Matrix = graphUtil.normalizeKernel(Matrix)
# start cross-validation for this t
for index_train, index_test in valid.k_fold_index(N, K=10):
y_train = graph_label[index_train, :]
y_test = graph_label[index_test, :]
n1 = len(index_train)
n2 = len(index_test)
model = pyGPs.GPC()
M1, M2 = graphUtil.formKernelMatrix(Matrix, index_train, index_test)
k = pyGPs.cov.Pre(M1, M2)
model.setPrior(kernel=k)
# gp
x_train = np.zeros((n1, 1))
x_test = np.zeros((n2, 1))
model.fit(x_train, y_train)
model.predict(x_test)
predictive_class = np.sign(model.ym)
# evaluation
acc = valid.ACC(predictive_class, y_test)
ACC.append(acc)
print 'Accuracy: ', np.round(np.mean(ACC),
2), '(' + str(np.round(np.std(ACC), 2)) + ')'
# State model with 10 classes
model = pyGPs.GPMC(10)
# Set data to model
model.setData(x,y)
# optimize default GPC model (see demo_GPC) for each binary classification problem,
# and decide label for test patterns of hand-writen digits
# prdictive_vote[i,j] is the probability of being class j for test pattern i
predictive_vote = model.optimizeAndPredict(xs)
predictive_class = np.argmax(predictive_vote, axis=1)
predictive_class = np.reshape(predictive_class, (predictive_class.shape[0],1))
# Accuracy of recognized digit
acc = valid.ACC(predictive_class, ys)
print "Accuracy of recognizing hand-writen digits:", round(acc,2)
#----------------------------------------------------------------------
# A bit more things you can do
#----------------------------------------------------------------------
# Just like we did for GP classification
# You can use specify the setting for all binary classificiation problem by:
m = pyGPs.mean.Zero()
k = pyGPs.cov.RBF()
model.setPrior(mean=m,kernel=k)
model.useInference("Laplace")
# Beside optimizeAndPredict(xs),
# there is also an option to predict without optimization
cv_run = 0
for x_train, x_test, y_train, y_test in valid.k_fold_validation(x, y, K):
print('Run:', cv_run)
# This is a binary classification problem
model = pyGPs.GPC()
# Since no prior knowldege, leave everything default
model.optimize(x_train, y_train)
# Predit
ymu, ys2, fmu, fs2, lp = model.predict(x_test, ys=y_test)
# ymu for classification is a continuous value over -1 to +1
# If you want predicting result to either one of the classes, take a sign of ymu.
ymu_class = np.sign(ymu)
# Evluation
acc = valid.ACC(ymu_class, y_test)
print(' accuracy =', round(acc, 2))
rmse = valid.RMSE(ymu_class, y_test)
print(' rmse =', round(rmse, 2))
ACC.append(acc)
RMSE.append(rmse)
# Toward next run
cv_run += 1
print('\nAccuracy: ', np.round(np.mean(ACC), 2),
'(' + str(np.round(np.std(ACC), 2)) + ')')
print('Root-Mean-Square Error: ', np.round(np.mean(RMSE), 2))
#----------------------------------------------------------------------
# Evaluation measures
