CoCalc Logo Icon
StoreFeaturesDocsShareSupportNewsAboutSign UpSign In
veeralakrishna

Real-time collaboration for Jupyter Notebooks, Linux Terminals, LaTeX, VS Code, R IDE, and more,
all in one place. Commercial Alternative to JupyterHub.

GitHub Repository: veeralakrishna/DataCamp-Project-Solutions-Python
 Path: blob/master/Who's Tweeting_ Trump or Trudeau_/datasets/helper_functions.py
Views: 1229
1
from matplotlib import pyplot as plt

2
import numpy as np

3
import itertools

4


5


6
def plot_confusion_matrix(cm, classes,

7
                          normalize=False,

8
                          title='Confusion matrix',

9
                          cmap=plt.cm.Blues,

10
                          figure=0):

11
    """

12
    See full source and example:

13
    http://scikit-learn.org/stable/auto_examples/model_selection/plot_confusion_matrix.html

14


15
    This function prints and plots the confusion matrix.

16
    Normalization can be applied by setting `normalize=True`.

17
    """

18
    plt.figure(figure)

19
    plt.imshow(cm, interpolation='nearest', cmap=cmap)

20
    plt.title(title)

21
    plt.colorbar()

22
    tick_marks = np.arange(len(classes))

23
    plt.xticks(tick_marks, classes, rotation=45)

24
    plt.yticks(tick_marks, classes)

25


26
    if normalize:

27
        cm = cm.astype('float') / cm.sum(axis=1)[:, np.newaxis]

28
        print("Normalized confusion matrix")

29
    else:

30
        print('Confusion matrix, without normalization')

31


32
    thresh = cm.max() / 2.

33
    for i, j in itertools.product(range(cm.shape[0]), range(cm.shape[1])):

34
        plt.text(j, i, cm[i, j],

35
                 horizontalalignment="center",

36
                 color="white" if cm[i, j] > thresh else "black")

37


38
    plt.tight_layout()

39
    plt.ylabel('True label')

40
    plt.xlabel('Predicted label')

41


42


43
def plot_and_return_top_features(classifier, vectorizer, top_features=20):

44
    """

45
    Plot the top features in a binary classification model and remove possible overlap.

46


47
    Adapted from https://medium.com/@aneesha/visualising-top-features-in-linear-svm-with-scikit-learn-and-matplotlib-3454ab18a14d

48
    and https://stackoverflow.com/a/26980472 by @kjam

49
    """

50
    class_labels = classifier.classes_

51
    feature_names = vectorizer.get_feature_names()

52
    topn_class1 = sorted(zip(classifier.coef_[0], feature_names))[:top_features]

53
    topn_class2 = sorted(zip(classifier.coef_[0], feature_names))[-top_features:]

54
    top_coefficients = np.hstack([topn_class1, topn_class2])

55
    if set(topn_class1).union(topn_class2):

56
        top_coefficients = topn_class1

57
        for ce in topn_class2:

58
            if ce not in topn_class1:

59
                top_coefficients.append(x)

60


61
    plt.figure(figsize=(15, 5))

62
    colors = ['red' if c < 0 else 'blue' for c in [tc[0] for tc in top_coefficients]]

63
    plt.bar(np.arange(len(top_coefficients)), [tc[0] for tc in top_coefficients], color=colors)

64
    plt.xticks(np.arange(len(top_coefficients)),

65
               [tc[1] for tc in top_coefficients], rotation=60, ha='right')

66
    plt.show()

67
    return top_coefficients

68


69