print("\n")
path = 'model.pth'
#torch.save(model.state_dict(), path)
elif args.mode == "topNsimilar":
path = 'model.pth'
state_dict = torch.load(path)
embedding = state_dict['input_embedding.weight'].to('cpu')
Words = getSimilarWords(embedding, args.word, args.N_words, idx2word, word2idx)
print(", ".join(words))
elif args.mode == "Analogy":
path = 'model.pth'
state_dict = torch.load(path)
embedding = state_dict['input_embedding.weight'].to('cpu')
words = getAnalogy(embedding, args.word1, args.word2, args.word3, 5, idx2word, word2idx)
print(words[1])
elif args.mode == "GetScore":
path = 'model.pth'
state_dict = torch.load(path)
embedding = state_dict['input_embedding.weight'].to('cpu')
print(getScore(embedding, args.word1, args.word2, idx2word, word2idx))
elif args.mode == "plot":
path = 'model.pth'
state_dict = torch.load(path)
embedding = state_dict['input_embedding.weight'].to('cpu')
plotData(embedding,idx2word)
for index in word_indices:
x[index] = 1
# ===========================================================
return x
# -------------------------- Testing Gaussian Kernel --------------------------------------
# Load from ex6data1
# You will have X, y as keys in the dict data
data = loadmat(os.path.join('Data', 'ex6data1.mat'))
X, y = data['X'], data['y'][:, 0]
# Plot training data
utils.plotData(X, y)
#pyplot.show()
# You should try to change the C value below and see how the decision
# boundary varies (e.g., try C = 1000)
C = 1
model = utils.svmTrain(X, y, C, utils.linearKernel, 1e-3, 20)
utils.visualizeBoundaryLinear(X, y, model)
#pyplot.show()
x1 = np.array([1, 2, 1])
x2 = np.array([0, 4, -1])
sigma = 2
sim = gaussianKernel(x1, x2, sigma)
testInputs = [inputs[i] for i in testSample]
testOutputs = [numericalOutputs[i] for i in testSample]
# print(trainInputs)
# print(testInputs)
trainInputs, testInputs = \
zScoreNormalization(trainInputs, testInputs)
# print(trainInputs)
# print(testInputs)
trainFeature1 = [sample[0] for sample in trainInputs]
trainFeature2 = [sample[1] for sample in trainInputs]
trainFeature3 = [sample[2] for sample in trainInputs]
trainFeature4 = [sample[3] for sample in trainInputs]
plotData(trainFeature1, trainFeature2, trainFeature3, trainFeature4,
"Train data")
# tool
classifier = linear_model.LogisticRegression(max_iter=1000)
classifier.fit(trainInputs, trainOutputs)
print("-----------tool-----------")
print("Real: " + str(testOutputs))
print("Comp: " + str(list(classifier.predict(testInputs))))
print("Accuracy: " +
str(accuracy(testOutputs, list(classifier.predict(testInputs)))))
# manual
classifierIrisSetosa = MyLogisticRegression()
classifierIrisVersicolor = MyLogisticRegression()
classifierIrisVirginica = MyLogisticRegression()
def main():
# Read training set
print
print 'Reading training set...'
train = utils.ReadFile('train.csv', 1)
print 'Finished reading...\n'
# Preliminary Statistics
print 'Preliminary Statistics:'
print np.shape(train)[0] - 1, 'people.', np.shape(train)[1] - 2, 'features.'
print (train[1:,1] == '1').sum(), 'survivors.', (train[1:,1] == '0').sum(), 'deceased.\n'
#Testing
id = 10
mask = train[1:,id] == ''
#print list(set(tmp))
#print train[1:,id]
#print mask.sum()
# Map string features to floats
print 'Mapping Features to Floats.\n'
dictN = {} # modified in call (useful for name feature)
dictC = {} # modified in call (useful for cabin feature)
dat, dictN, dictC = utils.mapToF(train[1:,:], 0, dictN, dictC)
# Class labels
lab = np.array([int(h) for h in train[1:,1]])
# Generate better model for missing Age feature
#means = np.zeros(len(dictN), dtype = np.float64)
#dat, means = utils.AgeModel(dat, dictN, means, 1)
mask = dat[:,2] != -1.0
dat2 = np.zeros((mask.sum(),9), dtype = np.float64)
tar2 = np.zeros(mask.sum(), dtype = np.float64)
dat, dat2, tar2 = utils.AgeModel2(dat, dat2, tar2, 1)
# testing
#mask = dat[:,2] == -1.0
#print mask.sum()
# Preliminary Plots
print 'Generating preliminary scatter plots of data.\n'
utils.PrelimPlots(dat, lab)
utils.AgePlots(dat)
#dat = utils.MeanNorm(dat)
# ML algorithms
print "Choosing best parameters for Random Forest algorithm:"
optim = TestRandForest(dat, lab)
# Plotting Learning Curve
print
print "Plotting the learning curve\n"
plotLearningCurve(dat, lab, optim)
# Where is algorithm failing?
print "Where is algorithm failing:\n"
whereFailing(dat, lab, optim)
# Read in test set
print "Reading in Test Set\n"
test = utils.ReadFile('test.csv', 0)
# Map to floats
testF, dictN, dictC = utils.mapToF(test[1:,:], 1, dictN, dictC)
# Make better prediction for missing Age Features
#testF, means = utils.AgeModel(testF, dictN, means, 0)
testF, dat2, tar2 = utils.AgeModel2(testF, dat2, tar2, 0)
# Generate scatter plot for test set
utils.plotData(testF, 0)
#testF = utils.MeanNorm(testF)
# Make prediction
print "Making Prediction\n"
clf = RandomForestClassifier(n_estimators = optim[0],
max_features = optim[1],
min_samples_split = 1)
clf = clf.fit(dat, lab)
pred = clf.predict(testF)
# Now output prediction
print "Outputting Prediction\n"
utils.OutputFile(pred, train[0,:2], test[1,0], 0)
print "Done"
data = pd.read_csv('./data/ex2data1.txt',
sep=',',
names=['test1', 'test2', 'exam'])
X = data.iloc[:, :2].values
y = data.iloc[:, 2].values
print(data.head())
# ==================== Part 1: Plotting ====================
# We start the exercise by first plotting the data to understand the
# the problem we are working with.
#
print(
'Plotting data with + indicating (y = 1) examples and o indicating (y = 0) examples.)'
)
plotData(X, y)
plt.legend(['Admitted', 'Not admitted'],
loc='upper right',
shadow=True,
fontsize='x-large',
numpoints=1)
plt.xlabel('Exam 1 score')
plt.ylabel('Exam 2 score')
# plt.show()
input("Program paused. Press Enter to continue...")
# ============ Part 2: Compute Cost and Gradient ============
# In this part of the exercise, you will implement the cost and gradient
# for logistic regression. You neeed to complete the code in
# costFunction.m
def main():
# Read training set
print
print 'Reading training set...'
train = utils.ReadFile('train.csv', 1)
print 'Finished reading...\n'
# Preliminary Statistics
print 'Preliminary Statistics:'
print np.shape(train)[0] - 1, 'people.', np.shape(
train)[1] - 2, 'features.'
print(train[1:, 1] == '1').sum(), 'survivors.', (
train[1:, 1] == '0').sum(), 'deceased.\n'
#Testing
id = 10
mask = train[1:, id] == ''
#print list(set(tmp))
#print train[1:,id]
#print mask.sum()
# Map string features to floats
print 'Mapping Features to Floats.\n'
dictN = {} # modified in call (useful for name feature)
dictC = {} # modified in call (useful for cabin feature)
dat, dictN, dictC = utils.mapToF(train[1:, :], 0, dictN, dictC)
# Class labels
lab = np.array([int(h) for h in train[1:, 1]])
# Generate better model for missing Age feature
#means = np.zeros(len(dictN), dtype = np.float64)
#dat, means = utils.AgeModel(dat, dictN, means, 1)
mask = dat[:, 2] != -1.0
dat2 = np.zeros((mask.sum(), 9), dtype=np.float64)
tar2 = np.zeros(mask.sum(), dtype=np.float64)
dat, dat2, tar2 = utils.AgeModel2(dat, dat2, tar2, 1)
# testing
#mask = dat[:,2] == -1.0
#print mask.sum()
# Preliminary Plots
print 'Generating preliminary scatter plots of data.\n'
utils.PrelimPlots(dat, lab)
utils.AgePlots(dat)
#dat = utils.MeanNorm(dat)
# ML algorithms
print "Choosing best parameters for Random Forest algorithm:"
optim = TestRandForest(dat, lab)
# Plotting Learning Curve
print
print "Plotting the learning curve\n"
plotLearningCurve(dat, lab, optim)
# Where is algorithm failing?
print "Where is algorithm failing:\n"
whereFailing(dat, lab, optim)
# Read in test set
print "Reading in Test Set\n"
test = utils.ReadFile('test.csv', 0)
# Map to floats
testF, dictN, dictC = utils.mapToF(test[1:, :], 1, dictN, dictC)
# Make better prediction for missing Age Features
#testF, means = utils.AgeModel(testF, dictN, means, 0)
testF, dat2, tar2 = utils.AgeModel2(testF, dat2, tar2, 0)
# Generate scatter plot for test set
utils.plotData(testF, 0)
#testF = utils.MeanNorm(testF)
# Make prediction
print "Making Prediction\n"
clf = RandomForestClassifier(n_estimators=optim[0],
max_features=optim[1],
min_samples_split=1)
clf = clf.fit(dat, lab)
pred = clf.predict(testF)
# Now output prediction
print "Outputting Prediction\n"
utils.OutputFile(pred, train[0, :2], test[1, 0], 0)
print "Done"
regressor2.fit(trainInputs, trainOutputs)
wPrim = [regressor2.intercept_, regressor2.coef_[0], regressor2.coef_[1]]
print("-----manual-----")
print("The learnt model: f(X,w) = " + str(wPrim[0]) + " + " +
str(wPrim[1]) + " * X1 + " + str(wPrim[2]) + " * X2")
# print("Real: " + str(testOutputs))
# print("Computed: " + str(list(regressor2.predict(testInputs))))
print("-----performance-----")
print("Prediction error (tool): ",
str(mean_squared_error(testOutputs, regressor2.predict(testInputs))))
print("Prediction error (manual): ",
str(meanSquareError(testOutputs, regressor2.predict(testInputs))))
plotDataHistogram(gdpData, 'GDP')
plotDataHistogram(freedomData, 'Freedom')
plotDataHistogram(outputs, 'Happiness score')
# for train and test data
plotData(gdpData, freedomData, outputs, w, "Train & test data")
# for train data
plotData(trainGdp, trainFreedom, trainOutputs, w,
"Train data and the learnt model")
# for test data
computedTestOutputs = regressor2.predict(testInputs)
plotData2(testGdp, testFreedom, testOutputs, computedTestOutputs,
"Computed(green) vs real(red) test data")
