19mach_learning.py

from sklearn import datasets
from sklearn.decomposition import PCA
from sklearn.model_selection import train_test_split
from sklearn.neighbors import KNeighborsClassifier
import matplotlib.pyplot as plt
import numpy as np

digits = datasets.load_digits()
X = digits["data"]
Y = digits["target"]

pca = PCA(n_components=2)
X_pca = pca.fit_transform(X)

train_ind, test_ind, = train_test_split(range(len(X)), test_size = 0.25)

X_train = X[train_ind]
Y_train = Y[train_ind]

X_test = X[test_ind]
Y_test = Y[test_ind]

classifier = KNeighborsClassifier(n_neighbors=20)
classifier.fit(X_train, Y_train)
predicted = classifier.predict(X_test)
labels = np.unique(Y)
colors = ["violet", "blue", "red"]

for color, label in zip(colors, labels):
    plt.plot(X_pca[train_ind][Y_train == label][:, 0], X_pca[train_ind][Y_train == label][:, 1], "o", c=color)

for color, label in zip(colors, labels):
    plt.plot(X_pca[test_ind][predicted == label][:, 0], X_pca[test_ind][predicted == label][:, 1], "x", c=color)

#plt.show()

predict = classifier.predict(X_test)

correct_num = len(list(filter(lambda el: el[0] == el[1], zip(Y_test, predicted))))
print("Total:")
print(correct_num/len(predicted))

for label in labels:
    print("Class {}:".format(label))
    y_for_label = list(filter(lambda x: x == label, Y_test))
    predicted_y_for_label = list(filter(lambda x: x == label, predicted))
    len_ = len(y_for_label)
    correct_ = len(list(filter(lambda el: el[0] == el[1], zip(y_for_label, predicted_y_for_label))))
    print(correct_ / len_)