svr.fit(X, Y)
t2 = time.time()
print(mem_us_svr)
svm_err = mean_squared_error(svr.predict(x), y)
svs = svr.best_estimator_.support_vectors_.shape[0]
print "SVM error on test set is {0}, number of support vectors is {1}, time {2}".format(
svm_err, svs, t2 - t1)
# Time
# Memory usage
# Number of SV
# Create Prediction
nPred = 1000
xPred = np.linspace(min(Xc), max(Xc), nPred).reshape((nPred, 1))
y_RVR_pred, var_RVR_pred = rvr.predict_dist(xPred)
#y_SVR_pred, var_SVR_pred = svr.predict_dist(xPred) NOT possible?
y_SVR_pred = svr.predict(xPred)
y_real = 10 * np.sinc(xPred)
# plot test vs predicted datax
plt.figure(figsize=(16, 10))
plt.plot(X[:, 0], Y, "k+", markersize=3, label="train data")
plt.plot(x[:, 0], y, "b+", markersize=3, label="test data")
plt.plot(xPred[:, 0], y_RVR_pred, "b", markersize=3, label="RVR prediction")
#plt.plot(xPred[:,0],y_RVR_pred + np.sqrt(var_RVR_pred),"c", markersize = 3, label = "y_hat +- std")
#plt.plot(xPred[:,0],y_RVR_pred - np.sqrt(var_RVR_pred),"c", markersize = 3)
plt.plot(rvr.relevant_vectors_,
Y[rvr.active_],
"co",
# parameters
n = 5000
# generate data set
np.random.seed(0)
Xc = np.ones([n, 1])
Xc[:, 0] = np.linspace(-5, 5, n)
Yc = 10 * np.sinc(Xc[:, 0]) + np.random.normal(0, 1, n)
X, x, Y, y = train_test_split(Xc, Yc, test_size=0.5, random_state=0)
# train rvr
rvm = RVR(gamma=1, kernel='rbf')
t1 = time.time()
rvm.fit(X, Y)
t2 = time.time()
y_hat, var = rvm.predict_dist(x)
rvm_err = mean_squared_error(y_hat, y)
rvs = np.sum(rvm.active_)
print("RVM error on test set is {0}, number of relevant vectors is {1}, time {2}".format(rvm_err, rvs, t2 - t1))
# train svr
svr = GridSearchCV(SVR(kernel='rbf', gamma=1), param_grid={'C': [0.001, 0.1, 1, 10, 100]}, cv=10)
t1 = time.time()
svr.fit(X, Y)
t2 = time.time()
svm_err = mean_squared_error(svr.predict(x), y)
svs = svr.best_estimator_.support_vectors_.shape[0]
print(print()
)
# # plot test vs predicted data
mem_us_svr = memory_usage((svr.fit,(X,Y)),interval=0.1)
#svr.fit(X,Y)
t2 = time.time()
minMem_svr = min(mem_us_svr)
maxMem_svr = max(mem_us_svr)
svm_err = mean_squared_error(svr.predict(x),y)
svs = svr.best_estimator_.support_vectors_.shape[0]
print "SVR -- NMSR {0}, # SV {1}, time {2}, min Memory {3}, max. Memory {4}".format(svm_err, svs, t2 - t1, minMem_svr, maxMem_svr)
# Time
# Memory usage
# Number of SV
# Create Prediction
dimTrain = Xc.shape
nPred = 1000
xPred = np.zeros((nPred,dimTrain[1]))
for i in range(dimTrain[1]):
xPred[:,i] = np.linspace(min(Xc[:,i]),max(Xc[:,i]),nPred).reshape((nPred,))
y_RVR_pred, var_RVR_pred = rvr.predict_dist(xPred)
y_SVR_pred = svr.predict(xPred)
print('program succesfully terminated')