randomforest.py

import pandas as pd
import numpy as np
from matplotlib import pyplot as plt
import utils

from sklearn.tree import DecisionTreeClassifier
from sklearn import tree

np.random.seed(0)

# Spam Email dataset
emails = np.array([
    [7,8,1],
    [3,2,0],
    [8,4,1],
    [2,6,0],
    [6,5,1],
    [9,6,1],
    [8,5,0],
    [7,1,0],
    [1,9,1],
    [4,7,0],
    [1,3,0],
    [3,10,1],
    [2,2,1],
    [9,3,0],
    [5,3,0],
    [10,1,0],
    [5,9,1],
    [10,8,1],
])
spam_dataset = pd.DataFrame(data=emails, columns=["Lottery", "Sale", "Spam"])
spam_dataset

# Data plot
features = spam_dataset[['Lottery', 'Sale']]
labels = spam_dataset['Spam']
utils.plot_points(features, labels)

# Decision Tree
decision_tree_classifier = DecisionTreeClassifier(random_state=0)
decision_tree_classifier.fit(features, labels)
decision_tree_classifier.score(features, labels)

# Draw decision tree
utils.display_tree(decision_tree_classifier)

# Decision tree as map
utils.plot_model(features, labels, decision_tree_classifier)

# Training a Random Forest
from sklearn.ensemble import RandomForestClassifier
random_forest_classifier = RandomForestClassifier(random_state=0, n_estimators=5, max_depth=1)
random_forest_classifier.fit(features, labels)
random_forest_classifier.score(features, labels)

# plot
utils.plot_model(features, labels, random_forest_classifier)

utils.plot_points(features, labels)

#
for dt in random_forest_classifier.estimators_:
    print("*"*30, "Estimator", "*"*30)
    tree.plot_tree(dt, rounded=True)
    plt.show()
    utils.plot_model(features, labels, dt)
    plt.show()
    
# Ada Boosting
from sklearn.ensemble import AdaBoostClassifier
# Set the random_state so that we always get the same results
adaboost_classifier = AdaBoostClassifier(random_state=0, n_estimators=6)
adaboost_classifier.fit(features, labels)
adaboost_classifier.score(features, labels)

utils.plot_model(features, labels, adaboost_classifier)