8/23/23, 11:40 PM 20MIS1025_DecisionTree.

ipynb - Colaboratory

Importing the libraries

from IPython.display import Image
%matplotlib inline

import os
from pathlib import Path
import pandas as pd
df = pd.read_csv("KDD_Train.csv")
df.replace(('normal','anomaly'), (0,1), inplace=True)

import warnings
warnings.filterwarnings('ignore')
import numpy as np

X = df.iloc[:,[4,9]].values
y=df.iloc[:, -1].values

print('Class labels:',np.unique(y))

Class labels: [0 1]

df.isnull().sum().sort_values(ascending=False).head()

duration 0
dst_host_count 0
srv_count 0
serror_rate 0
srv_serror_rate 0
dtype: int64

70% & 30% Splitting into Training & Testing Data

from sklearn.model_selection import train_test_split
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.3, random_state = 1, stratify = y)

print('Labels count in y', np.bincount(y))
print('Labels count in y_train', np.bincount(y_train))
print('Labels count in y_test', np.bincount(y_test))

Labels count in y [67343 58630]

Labels count in y_train [47140 41041]
Labels count in y_test [20203 17589]

Standardizing the features:

from sklearn.preprocessing import StandardScaler                                        #normalisation - converting values in the range 0
                                                                                        #all values are normalised
sc = StandardScaler()                                                                   #mean and standard deviation
sc.fit(X_train)                                                                         #fit against standard scalar                  con
X_train_std = sc.transform(X_train)                                                     #transformation of both training and test data is
X_test_std = sc.transform(X_test)

from matplotlib.colors import ListedColormap
import matplotlib.pyplot as plt

def plot_decision_regions(X, y, classifier, test_idx=None, resolution=0.02):

    # setup marker generator and color map
    markers = ('s', 'x', 'o', '^', 'v')
    colors = ('red', 'blue', 'lightgreen', 'gray', 'cyan')
    cmap = ListedColormap(colors[:len(np.unique(y))])

    # plot the decision surface
    x1_min, x1_max = X[:, 0].min() - 1, X[:, 0].max() + 1
    x2_min, x2_max = X[:, 1].min() - 1, X[:, 1].max() + 1
    xx1, xx2 = np.meshgrid(np.arange(x1_min, x1_max, resolution),
                           np.arange(x2_min, x2_max, resolution))
    Z = classifier.predict(np.array([xx1.ravel(), xx2.ravel()]).T)
    Z = Z.reshape(xx1.shape)
    plt.contourf(xx1, xx2, Z, alpha=0.3, cmap=cmap)
    plt.xlim(xx1.min(), xx1.max())
    plt.ylim(xx2.min(), xx2.max())

https://colab.research.google.com/drive/1FB1tRB0tXqyCLvBpmguU9gQbm0Ki-Plo#scrollTo=paC5OhBLFmoJ&printMode=true 1/4
8/23/23, 11:40 PM 20MIS1025_DecisionTree.ipynb - Colaboratory

    for idx, cl in enumerate(np.unique(y)):
        plt.scatter(x=X[y == cl, 0],
                    y=X[y == cl, 1],
                    alpha=0.8,
                    c=colors[idx],
                    marker=markers[idx],
                    label=cl,
                    edgecolor='black')

    # highlight test examples
    if test_idx:
        # plot all examples
        X_test, y_test = X[test_idx, :], y[test_idx]

        plt.scatter(X_test[:, 0],
                    X_test[:, 1],
                    c='',
                    edgecolor='black',
                    alpha=1.0,
                    linewidth=1,
                    marker='o',
                    s=100,
                    label='test set')

Decision tree learning

Building a decision tree

from sklearn.tree import DecisionTreeClassifier

tree_model = DecisionTreeClassifier(criterion='gini',max_depth=2,random_state=1)
tree_model.fit(X_train, y_train)

X_combined = np.vstack((X_train, X_test))
y_combined = np.hstack((y_train, y_test))
#plot_decision_regions(X_combined, y_combined,classifier=tree_model,test_idx=range(105, 150))

plt.xlabel('petal length [cm]')
plt.ylabel('petal width [cm]')
plt.legend(loc='upper left')
plt.tight_layout()
#plt.savefig('images/03_20.png', dpi=300)
plt.show()

WARNING:matplotlib.legend:No artists with labels found to put in legend. Note that a

https://colab.research.google.com/drive/1FB1tRB0tXqyCLvBpmguU9gQbm0Ki-Plo#scrollTo=paC5OhBLFmoJ&printMode=true 2/4
8/23/23, 11:40 PM 20MIS1025_DecisionTree.ipynb - Colaboratory
from sklearn import tree

tree.plot_tree(tree_model)
#plt.savefig('images/03_21_1.pdf')
plt.show()

!pip3 install pydotplus

Requirement already satisfied: pydotplus in /usr/local/lib/python3.10/dist-packages (2.0.2)

Requirement already satisfied: pyparsing>=2.0.1 in /usr/local/lib/python3.10/dist-packages (from pydotplus) (3.1.1)

!conda install python-graphviz

/bin/bash: line 1: conda: command not found

!pip install graphviz

Requirement already satisfied: graphviz in /usr/local/lib/python3.10/dist-packages (0.20.1)

!pip3 install pydotplus

Requirement already satisfied: pydotplus in /usr/local/lib/python3.10/dist-packages (2.0.2)

Requirement already satisfied: pyparsing>=2.0.1 in /usr/local/lib/python3.10/dist-packages (from pydotplus) (3.1.1)

from pydotplus import graph_from_dot_data
from sklearn.tree import export_graphviz

dot_data = export_graphviz(tree_model,
                           filled=True,
                           rounded=True,
                           class_names=['Setosa',
                                        'Versicolor',
                                        'Virginica'],
                           feature_names=['petal length',
                                          'petal width'],
                           out_file=None)
graph = graph_from_dot_data(dot_data)
graph.write_png('tree.png')

True

https://colab.research.google.com/drive/1FB1tRB0tXqyCLvBpmguU9gQbm0Ki-Plo#scrollTo=paC5OhBLFmoJ&printMode=true 3/4
8/23/23, 11:40 PM 20MIS1025_DecisionTree.ipynb - Colaboratory

check 0s completed at 11:36 PM

https://colab.research.google.com/drive/1FB1tRB0tXqyCLvBpmguU9gQbm0Ki-Plo#scrollTo=paC5OhBLFmoJ&printMode=true 4/4