def build_and_run_rvc(X_train_scaled, y_train, X_test_scaled, y_test):
'''
Takes: training and testing data
Returns: time to fit, time to predict, plus it prints
'''
# build RVC
rvc_model = RVC()
print("fitting RVM:")
start = time.time()
rvc_model.fit(X_train_scaled, y_train)
delta0 = time.time() - start
print("time to fit RVM: ", delta0)
start = time.time()
rvc_predict = rvc_model.predict(X_test_scaled)
delta1 = time.time() - start
print("time to predict with RVM: ", delta1)
# print parameters
print("RVM hyperparameters:")
print(rvc_model.get_params())
# evaluate RVC
print(helpers.confusion_matrix(y_test, rvc_predict))
print(classification_report(y_test, rvc_predict))
return delta0, delta1
def test_fit_two_classes(self):
"""Check that fitting with two classes works directly."""
clf = RVC()
X = np.array([[1, 2], [2, 1]])
y = np.array(['A', 'B'])
clf.fit(X, y)
np.testing.assert_array_equal(clf.classes_, np.array(['A', 'B']))
def test_fit_three_classes(self):
"""Check that fitting with three classes uses OneVSOne."""
clf = RVC()
X = np.array([[1, 2], [2, 1], [2, 2]])
y = np.array(['A', 'B', 'C'])
clf.fit(X, y)
self.assertIsInstance(clf.multi_, OneVsOneClassifier)
np.testing.assert_array_equal(clf.classes_, np.array(['A', 'B', 'C']))
def test_classification_three_classes(self):
"""Check classification works with three classes."""
iris = load_iris()
X = iris.data
y = iris.target
clf = RVC()
clf.fit(X, y)
self.assertGreater(clf.score(X, y), 0.95)
def test_predict_three_classes(self):
"""Check predict works with three classes."""
clf = RVC(kernel='linear')
X = np.array([[5, 5], [5, -5], [-5, 0]])
y = np.array(['A', 'B', 'C'])
clf.fit(X, y)
prediction = clf.predict(np.array([[10, 10]]))
np.testing.assert_array_equal(prediction, np.array(['A']))
def test_fit_one_class(self):
"""Check that fitting with only one class raises an exception."""
clf = RVC()
X = np.array([[1, 2], [2, 1]])
y = np.array(['A', 'A'])
try:
clf.fit(X, y)
except ValueError as error:
self.assertEqual(str(error), "Need 2 or more classes.")
else:
self.fail()
def test_predict_two_classes(self):
"""Check that predict works with two classes."""
clf = RVC(kernel='linear')
X = np.array([
[2, 1],
[1, 2],
])
y = np.array(['A', 'B'])
clf.fit(X, y)
prediction = clf.predict(np.array([[0, 3]]))
np.testing.assert_array_equal(prediction, np.array(['A']))
def test_fit_two_classes_imbalanced(self):
"""Check that fitting with two classes works with unequal samples."""
clf = RVC()
X = np.array([
[1, 2],
[1, 4],
[4, 2],
[2, 1],
[3, 1.5],
])
y = np.array(['A', 'A', 'B', 'B', 'B'])
clf.fit(X, y)
np.testing.assert_array_equal(clf.classes_, np.array(['A', 'B']))
def rvc_analysis(random_seed, save_path):
# Load the data
# TODO: change the path
save_path = os.path.join(save_path, 'random_seed_%03d' %random_seed)
print('Random seed: %03d' %random_seed)
# Load the saved validation dataset
project_ukbio_wd, project_data_ukbio, _ = get_paths(debug, dataset)
with open(os.path.join(save_path, 'splitted_dataset_%s.pickle' %dataset), 'rb') as handle:
splitted_dataset = pickle.load(handle)
# Train the model
model = RVC(kernel='linear')
model.fit(splitted_dataset['Xtrain_scaled'], splitted_dataset['Ytrain'])
# make cross validated predictions
print('Perform prediction in test data')
y_prediction_test = model.predict(splitted_dataset['Xtest_scaled'])
y_prediction_validation = model.predict(splitted_dataset['Xvalidate_scaled'])
# -----------------------------------------------------------------------------
# Do some statistics. Calculate the confusion matrix
# Test dataset
# Look at the confusion matrix for test data
class_name = np.array(['young', 'old', 'adult'], dtype='U10')
ax, cm_test = plot_confusion_matrix(splitted_dataset['Ytest'], y_prediction_test,
classes=class_name,
normalize=True)
# Look at accuracy
accuracy_test = accuracy_score(splitted_dataset['Ytest'], y_prediction_test)
plt.savefig(os.path.join(save_path, 'confusion_matrix_test_rvc.eps'))
# Predict on the validation dataset
ax, cm_validation = plot_confusion_matrix(splitted_dataset['Yvalidate'], y_prediction_validation,
classes=class_name,
normalize=True)
plt.savefig(os.path.join(save_path, 'confusion_matrix_validation_rvc.eps'))
# Look at accuracy
accuracy_val = accuracy_score(splitted_dataset['Yvalidate'],
y_prediction_validation)
plt.savefig(os.path.join(save_path, 'confusion_matrix_test_rvc.eps'))
return cm_test, cm_validation, accuracy_test, accuracy_val
def test_classification_two_classes(self):
"""Check classification works with two classes."""
iris = load_iris()
X = iris.data[:, 1:]
y = iris.target
# Only 2 classes needed
X = X[y != 0]
y = y[y != 0]
clf = RVC()
clf.fit(X, y)
self.assertGreater(clf.score(X, y), 0.95)
prob = clf.predict_proba(X[0, :])
p_target = np.array([[0.999, 5.538e-4]])
np.testing.assert_allclose(prob, p_target, rtol=1e-2, atol=1e-2)
def TrainMyRVM(XEstimate,
XValidate,
ClassLabelsEstimate,
ClassLabelsValidate,
Parameters=None):
training_labels = np.int8(np.zeros(ClassLabelsEstimate.shape[0]))
validate_labels = np.int8(np.zeros(ClassLabelsValidate.shape[0]))
for i in range(ClassLabelsEstimate.shape[0]):
training_labels[i] = np.where(ClassLabelsEstimate[i] == 1)[0]
for i in range(ClassLabelsValidate.shape[0]):
validate_labels[i] = np.where(ClassLabelsValidate[i] == 1)[0]
#get 4000 samples of training data
#this will get the indices and will give the data and labels the same indices
idx_training = np.random.choice(np.arange(len(training_labels)),
4000,
replace=False)
training_labels_sampled = training_labels[idx_training]
XEstimate_sampled = XEstimate[idx_training]
#get 1000 samples of training data
#this will get the indices and will give the data and labels the same indices
idx_validate = np.random.choice(np.arange(len(validate_labels)),
1000,
replace=False)
XValidate_sampled = XValidate[idx_validate]
ClassLabelsValidate_sampled = ClassLabelsValidate[idx_validate]
#initialize RVM with classification (RVC class)
rvm = RVC(kernel='rbf', n_iter=1, alpha=1.e-6, beta=1.e-6, verbose=True)
#fit RVM
rvm.fit(XEstimate_sampled, training_labels_sampled)
#predict and return an array of classes for each input
Yvalidate = rvm.predict(XValidate_sampled)
EstParameters = rvm
Rvectors = 1
return Yvalidate, EstParameters, Rvectors, ClassLabelsValidate_sampled, idx2
def RVM(X_hyper, Y_hyper, X_train, Y_train, X_validate, Y_validate, params):
clf = RVC(n_iter=100, tol=0.1)
start = time.clock()
X_train_reduced = X_train
X_validate_reduced = X_validate
train_size = params['train_size']
test_size = params['test_size']
train = params['train']
if train:
clf.fit(X_train_reduced[:train_size, :], Y_train[:train_size])
writeObj('rvm_model.pkl', clf)
Y_pred = clf.predict(X_validate_reduced[:test_size])
return Y_pred, clf
else:
clf = readObj('rvm_model.pkl')
Y_pred = clf.predict(X_validate_reduced[:test_size])
return Y_pred, clf
print "training took ", time.clock() - start, "s"
"""
Authors: Mrunmayee Deshpande, Lu Gan, Bruce Huang, Abhishek Venkataraman
"""
import timeit
from skrvm import RVC
import numpy as np
import os.path
import scipy.io
from import_data import import_data
## Set data path
parsed_data_path = 'parsed_data/'
[X, Y, valX, valY, testX, testY] = import_data(parsed_data_path)
scipy.io.savemat('train.mat', dict(X=X, Y=Y))
scipy.io.savemat('val.mat', dict(valX=valX, valY=valY))
scipy.io.savemat('test.mat', dict(testX=testX, testY=testY))
## Train a RVM
clf = RVC(verbose=True)
print(clf)
clf.fit(valX, valY)
clf.score(testX, testY)
#Reintegrate the validation set into the train/test sets for RVC
RXTrainHvM = HCvMCI[train_inds,]
RYTrainHvM = YHvM[train_inds]
RXTestHvM = HCvMCI[test_inds,]
RYTestHvM = YHvM[test_inds]
#resampling w/ SMOTE to account for uneven sampling
[XTrainResHvM,YTrainResHvM] = SMOTE(random_state = 100,k_neighbors = 3).fit_resample(RXTrainHvM,RYTrainHvM)
[XTestResHvM,YTestResHvM] = SMOTE(random_state = 100,k_neighbors = 3).fit_resample(RXTestHvM,RYTestHvM)
from skrvm import RVC
RVCMod = RVC(kernel = 'linear',
verbose = True)
RVCMod.fit(XTrainResHvm,YTrainResHvM)
#create feature importance evaluation function
def RVMFeatImp(RVs):
NumRVs = RVs.shape[0]
SumD = 0
for RVNum in range(1,NumRVs):
d1 = RVs[RVNum-1,]
d2 = sum(numpy.ndarray.flatten(
RVs[numpy.int8(
numpy.setdiff1d(numpy.linspace(0,NumRVs-1,NumRVs),RVNum))]))
SumD = SumD + (d1/d2)
SumD = SumD/NumRVs
return SumD
def TrainMyClassifier(XEstimate, ClassLabels, XValidate, Parameters):
# RVM
if Parameters['algorithm'] == 'RVM':
Parameters = Parameters['parameters']
clf = RVC(alpha=Parameters.get('alpha'),
beta=Parameters.get('beta'),
n_iter=Parameters.get('n_iter'))
clf.fit(XEstimate, ClassLabels)
if np.shape(clf.classes_)[0] == 2:
Yvalidate = clf.predict_proba(XValidate)
else:
Yvalidate = predict_proba(clf, XValidate)
EstParameters = get_params(clf)
return Yvalidate, EstParameters
#SVM
elif Parameters['algorithm'] == 'SVM':
svc = get_svc(Parameters)
svc_train(svc, XEstimate, ClassLabels)
prob = svc_probability(svc, XValidate)
EstParameters = svc_get_para(svc)
prob_std = np.ndarray.std(prob, axis=1)[:, np.newaxis]
sigmoid = 1 - expit(prob_std)
Yvalidate = np.concatenate([prob, sigmoid], axis=1)
Yvalidate = Yvalidate / np.repeat(
(sigmoid + 1), axis=1, repeats=len(svc.classes_) + 1)
return Yvalidate, EstParameters
#GPR
elif Parameters["algorithm"] == "GPR":
# get the classes from the labels
classes = np.unique(ClassLabels, axis=0)
sorted(classes, reverse=True)
num_class = len(classes)
# get data and label based on classes
data = []
for cla in classes:
data.append(XEstimate[ClassLabels == cla])
target = []
for cla in classes:
target.append(ClassLabels[ClassLabels == cla])
# put data and label into a matrix, so that we could do a easier calculation for probability
# the following calculation is all based on the matrix
data_matrix = []
for i in range(num_class - 1):
data_matrix.append([])
for j in range(num_class - 1):
data_matrix[i].append(None)
target_matrix = []
for i in range(num_class - 1):
target_matrix.append([])
for j in range(num_class - 1):
target_matrix[i].append(None)
for i in range(num_class - 1):
for j in range(i, num_class - 1):
data_matrix[i][j] = np.concatenate([data[i], data[j + 1]],
axis=0)
target_matrix[i][j] = np.concatenate(
[target[i], target[j + 1]], axis=0)
classifier_matrix = []
for i in range(num_class - 1):
classifier_matrix.append([])
for j in range(num_class - 1):
classifier_matrix[i].append(None)
for i in range(num_class - 1):
for j in range(i, num_class - 1):
gpc_classifier = GaussianProcessClassifier(
kernel=Parameters["parameters"]["kernel"],
optimizer=Parameters["parameters"]["optimizer"],
n_restarts_optimizer=Parameters["parameters"]
["n_restarts_optimizer"],
max_iter_predict=Parameters["parameters"]
["max_iter_predict"],
warm_start=Parameters["parameters"]["warm_start"],
copy_X_train=Parameters["parameters"]["copy_X_train"],
random_state=Parameters["parameters"]["random_state"],
multi_class="one_vs_rest",
n_jobs=Parameters["parameters"]["n_jobs"])
gpc_classifier.fit(data_matrix[i][j], target_matrix[i][j])
classifier_matrix[i][j] = gpc_classifier
out_matrix = []
for i in range(num_class - 1):
out_matrix.append([])
for j in range(num_class - 1):
out_matrix[i].append(None)
for i in range(num_class - 1):
for j in range(i, num_class - 1):
out_matrix[i][j] = classifier_matrix[i][j].predict_proba(
XValidate)
# calculate the whole prediction prob
val_shape = XValidate.shape[0]
predict_prob_list = []
for i in range(num_class):
predict_prob_list.append(np.zeros(shape=[val_shape, 1]))
for i in range(num_class - 1):
for j in range(i, num_class - 1):
predict_prob_list[i] += out_matrix[i][j][:,
0][:, np.newaxis] / (
num_class * 2)
predict_prob_list[
j +
1] += out_matrix[i][j][:, 1][:,
np.newaxis] / (num_class * 2)
# get the result of num_class probability
result = np.concatenate(predict_prob_list, axis=1)
# calculate the probability for the one more class
std = np.std(result, axis=1)[:, np.newaxis]
other_prob = np.exp(-std) / (1 + np.exp(std * 5))
result = np.concatenate([result, other_prob], axis=1)
result = result / np.repeat(
(other_prob + 1), axis=1, repeats=num_class + 1)
# put all the parameters into a dict
estParameters = {}
estParameters["class_num"] = num_class
estParameters["parameters"] = []
for i in range(num_class - 1):
for j in range(i, num_class - 1):
estParameters["parameters"].append({
"log_marginal_likelihood_value_":
classifier_matrix[i][j].log_marginal_likelihood_value_,
"classes_":
classifier_matrix[i][j].classes_,
"n_classes_":
classifier_matrix[i][j].n_classes_,
"base_estimator_":
classifier_matrix[i][j].base_estimator_
})
return result, estParameters
from skrvm import RVC
from sklearn.model_selection import cross_val_score
from sklearn.model_selection import train_test_split
def load_data():
f = open('F:/importantfilecopy/tandomrepeat/2019.121test/5results', 'r')
# f = open('F:/importantfilecopy/tandomrepeat/3/1x/total.txt', 'r')
data_set = []
label_set = []
for line in f:
line = line.strip().split("\t")
line_list = line[0].strip().split(" ")
data_set.append(line_list)
label_set.append(line[1])
return data_set, label_set
data, label = load_data()
X_train, X_test, y_train, y_test = train_test_split(data,
label,
test_size=0.3,
random_state=0)
clf = RVC()
# clf.fit(rvm_data, rvm_target)
# print(clf.score(rvm_data, rvm_target))
clf.fit(X_train, y_train)
scoring = 'accuracy'
scores = cross_val_score(clf, X_test, y_test, cv=7)
print(scores.mean())
p1_classes = full_class_array[0:690]
#normalize
p1_normal_data = preprocessing.scale(p1_data) #normalize
#PCA
pca = PCA(10, svd_solver='auto')
pca.fit(p1_normal_data)
p1_pca_data = pca.transform(p1_normal_data) #transform data to xx components
p1_pca_data
#classes_numbered, class_numbers_names = class2numbers(p1_classes)
## RVM classification
clf1 = RVC()
clf1.fit(p1_pca_data, p1_classes)
pred = clf1.predict(p1_pca_data)
correct = 0
for i in range(np.size(pred, 0)):
if pred[i] == p1_classes[i]:
correct += 1
clf1
params = clf1.get_params
params.alpha_
clf1.predict_proba
from skrvm import RVC
from sklearn.datasets import load_iris
if __name__ == "__main__":
clf = RVC()
clf.verbose = True # Print iteration, alpha, beta, gamma, m, Relevance vectors
data = load_iris()
trainData = data.data
trainTargets = data.target
print(clf.fit(trainData, trainTargets))
#print(clf.score(trainData, trainTargets))
YTrainHvM = YHvM[TrainInds]
XValHvM = HCvMCI[ValInds, ]
YValHvM = YHvM[ValInds]
XTestHvM = HCvMCI[TestInds, ]
YTestHvM = YHvM[TestInds]
#Running RVC - HCvMCI
RXTrainHvM = HCvMCI[train_inds, ]
RYTrainHvM = YHvM[train_inds]
RXTestHvM = HCvMCI[test_inds, ]
RYTestHvM = YHvM[test_inds]
from skrvm import RVC
RVCMod = RVC(kernel='linear', verbose=True)
RVCMod.fit(RXTrainHvM, RYTrainHvM)
def RVMFeatImp(RVs):
NumRVs = RVs.shape[0]
SumD = 0
for RVNum in range(1, NumRVs):
d1 = RVs[RVNum - 1, ]
d2 = sum(
np.ndarray.flatten(RVs[np.int8(
np.setdiff1d(np.linspace(0, NumRVs - 1, NumRVs), RVNum))]))
SumD = SumD + (d1 / d2)
SumD = SumD / NumRVs
return SumD
New[i, :] = n.params #mdl = smt.AR(d_tot[100,:]).fit(maxlag=30, ic='aic', trend='nc') #est_order=smt.AR(d_tot[1,:]).select_order(maxlag=30, ic='aic', trend='nc') #print(est_order) #p_orders=np.zeros([1380,1]) #for i in range(1380): # X=AR(d_tot[i,:]) # n=X.fit(maxlag=4,ic='aic') # p_orders[i,0]=len(n.params) #np.mean(p_orders) #plt.scatter(New[:,1],New[:,3]) An_lab = Animal_label(tot) #shuffle data to take them in random order indx = random.sample(range(10350), 10350) a = New[indx, :] b = An_lab[indx] #define trainning and test set x_train = a[0:8000, :] y_train = b[0:8000] x_test = a[8001:, :] y_test = b[8001:] #SVM with 600 out of 690 samples as training data from skrvm import RVC clf1 = RVC(kernel='rbf') clf1.fit(x_train, y_train) clf1.score(x_test, y_test)
feature_scaler = StandardScaler()
X_train = feature_scaler.fit_transform(X_train)
X_test = feature_scaler.transform(X_test)
####################################################################################################################
#CROSS VALIDATION SPLIT IN K-folds
######################################################################################################################
kf = KFold(n_splits=5)
kf.get_n_splits(X_train,X_test)
print(kf)
#CREATE WIDTHS FOR GRID SEARCH
width1= np.linspace(-5,4,10)
width=10**width1
#create matrix to input the scores and widths of each iteration
score_width=np.ones([len(width),2])
##################
#FIRST for loop gives the values for different widths, SECOND for loop does Kfold_validation
for i in range(len(width)):
score=0
for train_index, test_index in kf.split(X_train):
print("TRAIN:", train_index, "TEST:", test_index)
X_train1, X_test1 = X_train[train_index], X_train[test_index]
y_train1, y_test1 =y_train[train_index], y_train[test_index]
clf1=RVC(kernel='rbf',coef1=width[i])
clf1.fit(X_train1,y_train1)
score=score+clf1.score(X_test1,y_test1)
score_width[i,0]=score
score_width[i,1]=width[i]
#########################################################################################################
idx=np.argmax(score_width[:,0])
best_width=score_width[idx,1]
#!/usr/bin/env python3 # -*- coding: utf-8 -*- """ Created on Thu Sep 20 21:26:07 2018 @author: pan """ from skrvm import RVC from sklearn.datasets import load_iris clf = RVC() iris = load_iris() clf.fit(iris.data, iris.target) #RVC(alpha=1e-06, beta=1e-06, beta_fixed=False, bias_used=True, coef0=0.0, #coef1=None, degree=3, kernel='rbf', n_iter=3000, n_iter_posterior=50, #threshold_alpha=1000000000.0, tol=0.001, verbose=False) clf.score(iris.data, iris.target)
X = full_normPCA123_array[train_indicies]
Y = full_isAnimal_array[train_indicies]
params = rvc_param_selection(X, Y, 5)
#
#params = rvc_param_selection2(X,Y,5)
####################TO TEST
test_err = np.zeros((15))
train_err = np.zeros((15))
coef = np.multiply(
[1000, 500, 200, 150, 100, 70, 60, 50, 40, 30, 20, 10, 5, 2, 1], 1e-6)
for i in range(15):
#coef[i] = 5e-5-(i+1)*1e-5
clf = RVC(kernel="rbf", coef1=coef[i]) # coef1: 1=46 0.1same
clf.fit(full_normPCA123_array[train_indicies],
full_subClass_array[train_indicies])
train_err[i] = clf.score(full_normPCA123_array[train_indicies],
full_subClass_array[train_indicies])
test_err[i] = clf.score(full_normPCA123_array[test_indicies],
full_subClass_array[test_indicies])
print(coef[i])
print(train_err[i])
print(test_err[i])
print("\n\n")
#####################################################################################################################
rvm_tested_coef1 = np.load("rvm_tested_coef1.npy")
rvm_tested_coef1_again = np.load("rvm_tested_coef1_again.npy")
train1 = np.load("rvm_tested_coef1_trainerr.npy")
train2 = np.load("rvm_tested_coef1_trainerr_again.npy")
pd.set_option('display.max_columns', 100)
pd.set_option('display.notebook_repr_html', True)
import seaborn as sns
sns.set_style("whitegrid")
sns.set_context("poster")
# RVM
from skrvm import RVC
clf = RVC()
### read data from single csv file and transform to a list
dir = '/Users/weng/Downloads/'
# raw = pd.read_csv(os.path.join(dir, 'test.txt'), header=0, delimiter=';')
raw_list = pd.read_csv(os.path.join(dir, 'test.txt'), sep="\t", header=None)
# array format is needed for further processing (df -> list -> matrix(array) )
raw_X = raw_list[0].values.tolist()
raw_X = np.asarray(raw_X)
raw_y = raw_list[1].values.tolist()
raw_y = np.asarray(raw_y)
print "Read %d rows of data\n" % len(raw)
clf.fit(raw_X, raw_y)
from sklearn.datasets import load_iris
iris = load_iris()
iris.data
iris.target
clf.fit(iris.data, iris.target)
clf.score(iris.data, iris.target)
