def Highlighting_Test_Data_Points():
from mlxtend.plotting import plot_decision_regions
from mlxtend.preprocessing import shuffle_arrays_unison
import matplotlib.pyplot as plt
from sklearn import datasets
from sklearn.svm import SVC
# Loading some example data
iris = datasets.load_iris()
data = pd.read_csv('2clstrain1200.csv', header=None)
X, y = data.iloc[:, 0:2].values, data.iloc[:, 2].values
X = X.astype(np.integer)
y = y.astype(np.integer)
X, y = shuffle_arrays_unison(arrays=[X, y], random_seed=3)
X_train, y_train = X[:700], y[:700]
X_test, y_test = X[700:], y[700:]
# Training a classifier
svm = SVC(C=0.5, kernel='linear')
svm.fit(X_train, y_train)
# Plotting decision regions
plot_decision_regions(X, y, clf=svm, legend=2, X_highlight=X_test)
# Adding axes annotations
plt.xlabel('')
plt.ylabel('')
plt.title('SVM on Iris')
plt.show()
def main(self):
print(__doc__)
#creating object instances
obj = sat()
#extracting features and output
x, y = obj.load_data("gpa.csv")
#splitting the data
x_train, x_test, y_train, y_test = obj.split(x, y)
#scaling the data
#x_train,x_test,y_train,y_test = obj.scale(x_train,x_test,y_train,y_test)
#missing value imputation
x_train = obj.missing_val(x_train)
x_test = obj.missing_val(x_test)
y_train = obj.missing_val(y_train)
y_test = obj.missing_val(y_test)
#generating classifier
clf = obj.classifier()
#fitting the features into the model
clf.fit(x_train, y_train)
#plotting training set
obj.plot(clf, x_train, y_train, "orange", "blue",
"sat score (Training set)", "GPA", "SAT SCORE")
#plotting the testing set
obj.plot(clf, x_test, y_test, "orange", "blue",
"sat score (Testing set)", "GPA", "SAT SCORE")
#saving classifier
obj.save_classifier(clf, "sat_score.pkl", "wb")
#loading the data
clf = obj.load_classifier("sat_score.pkl", "rb")
x, y = shuffle_arrays_unison(arrays=[x, y], random_seed=5)
plot_learning_curves(x_train, y_train, x, y, clf)
plt.show()
def test_shuffle_arrays_unison():
X1 = np.array([[1, 2, 3], [4, 5, 6], [7, 8, 9]])
y1 = np.array([1, 2, 3])
X2, y2 = shuffle_arrays_unison(arrays=[X1, y1], random_seed=3)
assert (X2.all() == np.array([[4, 5, 6], [1, 2, 3], [7, 8, 9]]).all())
assert (y2.all() == np.array([2, 1, 3]).all())
def test_shuffle_arrays_unison():
X1 = np.array([[1, 2, 3], [4, 5, 6], [7, 8, 9]])
y1 = np.array([1, 2, 3])
X2, y2 = shuffle_arrays_unison(arrays=[X1, y1], random_seed=3)
assert(X2.all() == np.array([[4, 5, 6], [1, 2, 3], [7, 8, 9]]).all())
assert(y2.all() == np.array([2, 1, 3]).all())
def batch_generator(file_siize):
if 1000 % file_size != 0:
print('wrong size!')
os._exit()
global batch_count
file_num = int(batch_count % 11)
batch_count += 1
filenames = glob('data/radarTrend.train.*.input.npy')
filenames.sort()
x_train = np.load(filenames[file_num])
filenames = glob('data/radarTrend.train.*.label.npy')
filenames.sort()
y_train = np.load(filenames[file_num])
x_train, y_train = shuffle_arrays_unison(arrays=[x_train, y_train])
x_train = np.split(x_train, file_size)
y_train = np.split(y_train, file_size)
x_train = np.array(x_train)
y_train = np.array(y_train)
file_num = np.random.choice(10)
filenames = glob('data/radarTrend.test.*.input.npy')
x_valid = np.load(filenames[file_num])
filenames = glob('data/radarTrend.test.*label.npy')
y_valid = np.load(filenames[file_num])
x_valid, y_valid = shuffle_arrays_unison(arrays=[x_valid, y_valid])
x_valid = np.split(x_valid, file_size)
y_valid = np.split(y_valid, file_size)
x_valid = np.array(x_valid)
y_valid = np.array(y_valid)
return x_train, y_train, x_valid, y_valid
def Customizing_the_plotting_style():
from mlxtend.plotting import plot_decision_regions
from mlxtend.preprocessing import shuffle_arrays_unison
import matplotlib.pyplot as plt
from sklearn.svm import SVC
# Loading some example data
data = pd.read_csv('2clstrain1200.csv', header=None)
X, y = data.iloc[:, 0:2].values, data.iloc[:, 2].values
X = X.astype(np.integer)
y = y.astype(np.integer)
X, y = shuffle_arrays_unison(arrays=[X, y], random_seed=3)
X_train, y_train = X[:700], y[:700]
X_test, y_test = X[700:], y[700:]
# Training a classifier
svm = SVC(C=0.5, kernel='linear')
svm.fit(X_train, y_train)
# Specify keyword arguments to be passed to underlying plotting functions
scatter_kwargs = {'s': 120, 'edgecolor': None, 'alpha': 0.7}
contourf_kwargs = {'alpha': 0.2}
scatter_highlight_kwargs = {'s': 120, 'label': 'Test data', 'alpha': 0.7}
# Plotting decision regions
plot_decision_regions(X,
y,
clf=svm,
legend=2,
X_highlight=X_test,
scatter_kwargs=scatter_kwargs,
contourf_kwargs=contourf_kwargs,
scatter_highlight_kwargs=scatter_highlight_kwargs)
# Adding axes annotations
plt.xlabel('')
plt.ylabel('')
plt.title('SVM on Iris')
plt.show()
# dropout function
def dropout(a, prob):
shape = a.shape[0]
vec = np.random.choice([0,1], size = (shape,1), p = [prob, 1-prob])
return vec * a
#load dataset
iris = datasets.load_iris()
data = iris.data
actual = iris.target
# scramble arrays
data = inputn(data)
actual = inputn(actual)
# shuffle the arrays the same
data, actual = shuffle_arrays_unison(arrays=[data, actual], random_seed=3)
# initialize weights, divide by sqrt(4), to prevent explosive gradients
syn0 = np.random.randn(4,4) / 2
syn1 = np.random.randn(3,4) /2
bias0 = np.random.randn(4,1) /2
bias1 = np.random.randn(3,1) /2
# stores array of
xarr = []
yarr = []
rate_syn = 0.1
rate_bias = 0.1
for yo in range(200):
# Load the dataset - This dataset has 60 columns plus the decision variable 'R' and 'M'. It has a total of 208 rows.
X = np.loadtxt('sonar.txt', usecols=range(0, 60), delimiter=',')
labels = np.genfromtxt('sonar.txt', delimiter=',', usecols=-1, dtype=str)
# Traverse through the array Labels and replace 'R' with 1 and 'M' with 0.
for i, label in enumerate(labels):
if label == 'R':
labels[i] = 1
else:
labels[i] = 0
y = np.array(labels)
# Shuffle the data so that the test data towards the end does not end up with only 0's
from mlxtend.preprocessing import shuffle_arrays_unison
X, y = shuffle_arrays_unison(arrays=[X, y], random_seed=3)
# Convert y from an array of strings to array of integers
y = y.astype(np.int)
# Initialize the MLP class and fit the multi-layer perceptron model on X and y data. X and y is split implicitly inside the
# MLP class into training and test data
# Then predict the y values for X_test and store it in y_predicted
# The default number of iterations is 100, but can be changed by passing a value to n_iters. Similarly, learning rate is
# 0.01, but can be changed. In each iteration, we are printing the accuracy metrics for the predicted values of the training
# data.
mlp = MLP()
mlp.fit(X, y)
y_predicted = mlp.predict()
# An accuracy of 70% has been gained on test data after 100 epochs.
# mnist.py
from mlxtend.data import mnist_data
from mlxtend.classifier import MultiLayerPerceptron as MLP
from mlxtend.preprocessing import shuffle_arrays_unison
X, y = mnist_data()
X, y = shuffle_arrays_unison((X, y), random_seed=1)
X_train, y_train = X[:500], y[:500]
X_test, y_test = X[500:], y[500:]
import matplotlib.pyplot as plt
def plot_digit(X, y, idx):
img = X[idx].reshape(28, 28)
plt.imshow(img, cmap='Greys', interpolation='nearest')
plt.title('true label: %d' % y[idx])
plt.show()
plot_digit(X, y, 3500)
from mlxtend.preprocessing import standardize
X_train_std, params = standardize(X_train,
columns=range(X_train.shape[1]),
return_params=True)
X_test_std = standardize(X_test, columns=range(X_test.shape[1]), params=params)
nn1 = MLP(hidden_layers=[150],
l2=0.00,
l1=0.0,
epochs=100,
# In[47]:
plt.plot(range(len(nn2.cost_)), nn2.cost_)
plt.ylabel('Cost')
plt.xlabel('Epochs')
plt.show()
# ## Example 2 - Classifying Handwritten Digits from a 10% MNIST Subset
# In[53]:
from mlxtend.data import mnist_data
from mlxtend.preprocessing import shuffle_arrays_unison
X, y = mnist_data()
X, y = shuffle_arrays_unison((X, y), random_seed=1)
X_train, y_train = X[:500], y[:500]
X_test, y_test = X[500:], y[500:]
# Visualize a sample from the MNIST dataset to check if it was loaded correctly:
# In[55]:
import matplotlib.pyplot as plt
def plot_digit(X, y, idx):
img = X[idx].reshape(28, 28)
plt.imshow(img, cmap='Greys', interpolation='nearest')
plt.title('true label: %d' % y[idx])
plt.show()
# TODO error from mlxtend.evaluate import plot_learning_curves import matplotlib.pyplot as plt from mlxtend.data import iris_data from mlxtend.preprocessing import shuffle_arrays_unison from sklearn.neighbors import KNeighborsClassifier # Loading some example data X, y = iris_data() X, y = shuffle_arrays_unison(arrays=[X, y]) X_train, X_test = X[:100], X[100:] y_train, y_test = y[:100], y[100:] clf = KNeighborsClassifier(n_neighbors=5) plot_learning_curves(X_train, y_train, X_test, y_test, clf) plt.show()
bias0 = pickle.load(b0)
bias1 = pickle.load(b1)
# load MNIST dataset
mnist = fetch_mldata("MNIST original")
# learning rates for weights and bias
rate_syn = .0001
rate_bias = .0001
# data set
xarr = mnist.data
yarr = mnist.target
# shuffle arrays
xarr, yarr = shuffle_arrays_unison(arrays=[xarr, yarr], random_seed=4)
# only ujse first 1000 samples for training
xarr = xarr[:1000]
yarr = yarr[:1000]
# graphing variables
xbar = []
ybar = []
correct = 0
loss = 0
# training loop
for epoch in range(10):
for image in range(1000):
# convolution operation with max pooling
input = layer.convert_to_2d_image(xarr[image])
plt.show()
plt.plot(range(len(nn2.cost_)), nn2.cost_)
plt.ylabel('Cost')
plt.xlabel('Epochs')
plt.show()
# Classify handwritten digits from a 10% mnist subset
from mlxtend.data import mnist_data
from mlxtend.preprocessing import shuffle_arrays_unison, standardize
X, y = mnist_data()
X, y = shuffle_arrays_unison((X, y), random_seed=1)
X_train, y_train = X[:500], y[:500]
X_test, y_test = X[500:], y[500:]
def plot_digit(X, y, idx):
img = X[idx].reshape(28, 28) # 784 => 28 * 28
plt.imshow(img, cmap='Greys', interpolation='nearest')
plt.title('true label: %d' % y[idx])
plt.show()
plot_digit(X, y, 3500)
X_train_std, params = standardize(X_train, columns=range(X_train.shape[1]), return_params=True)
X_test_std = standardize(X_test, columns=range(X_test.shape[1]), params=params)
nn1 = MLP(hidden_layers=[150],