Dataset splitting for machine learning

Dataset splitting for machine learning

import numpy as np
import matplotlib.pyplot as plt
import pandas as pd

%matplotlib inline
%config InlineBackend.figure_format='retina'

# Create a dataset with 3 features and 1000 samples for a classification problem

df = pd.DataFrame(np.random.randn(1000, 3), columns=['A', 'B', 'C'])
df['D'] = np.random.randint(0, 2, 1000)

df

	A	B	C	D
0	0.112981	-0.033889	0.348708	1
1	1.821601	1.267423	0.274915	1
2	0.148040	-1.412712	-0.271345	1
3	0.004766	-1.209944	0.122512	1
4	0.854442	-0.559497	-0.605376	1
...	...	...	...	...
995	0.161444	-0.969567	-2.079664	0
996	1.615938	0.219218	-0.322223	0
997	0.501970	-0.874313	1.571102	1
998	0.386063	-1.481215	-1.974313	1
999	0.468528	1.060850	0.612252	1

1000 rows × 4 columns

X, y = df[['A', 'B', 'C']], df['D']

train_X, test_X = X[:800], X[800:]
train_y, test_y = y[:800], y[800:]

from sklearn.tree import DecisionTreeClassifier

dt = DecisionTreeClassifier(max_depth=2)
dt.fit(train_X, train_y)

DecisionTreeClassifier(max_depth=2)

In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook.
On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.

from sklearn import tree
tree.plot_tree(dt)

[Text(0.5, 0.8333333333333334, 'X[0] <= -0.17\ngini = 0.5\nsamples = 800\nvalue = [400, 400]'),
 Text(0.25, 0.5, 'X[0] <= -0.307\ngini = 0.495\nsamples = 348\nvalue = [157, 191]'),
 Text(0.125, 0.16666666666666666, 'gini = 0.499\nsamples = 316\nvalue = [153, 163]'),
 Text(0.375, 0.16666666666666666, 'gini = 0.219\nsamples = 32\nvalue = [4, 28]'),
 Text(0.75, 0.5, 'X[2] <= 1.656\ngini = 0.497\nsamples = 452\nvalue = [243, 209]'),
 Text(0.625, 0.16666666666666666, 'gini = 0.495\nsamples = 434\nvalue = [239, 195]'),
 Text(0.875, 0.16666666666666666, 'gini = 0.346\nsamples = 18\nvalue = [4, 14]')]

y_hat = dt.predict(test_X)
pd.DataFrame({"y_true": test_y, "y_hat": y_hat})

	y_true	y_hat
800	0	1
801	0	0
802	1	1
803	1	1
804	0	1
...	...	...
995	0	1
996	0	0
997	1	0
998	1	1
999	1	1

200 rows × 2 columns

from sklearn.tree import DecisionTreeClassifier

train_scores = {}
test_scores = {}

for i in range(1, 20):
    dt = DecisionTreeClassifier(max_depth=i)
    dt.fit(train_X, train_y)
    train_scores[i] = dt.score(train_X, train_y)
    
    test_scores[i] = dt.score(test_X, test_y)

scores_df = pd.DataFrame({'train': train_scores, 'test': test_scores})
scores_df.plot()

<AxesSubplot:>