Chapter 8: Study of Support Vector Machines (SVMs)

import pandas as pd

from sklearn.datasets import load_iris

iris = load_iris()

iris.feature_names

ds = pd.DataFrame(iris.data, columns=iris.feature_names)

ds.head()

# Appending one column as target column

ds[‘target’] = iris.target

ds.head()

# 0 ----> Setosa 1-->Versicolor 2---> Virginica

iris.target_names

array([‘setosa’, ‘versicolor’, ‘virginica’], dtype=‘<U10’)

#Want to see the number of rows for each flower type

ds[ds.target==2]

…

…

…

…

…

# Want to append flower name column ---> based on index (0, 1, 2) i.e Setosa and all

ds[‘fname’] = ds.target.apply(lambda x: iris.target_names[x])

ds.head()

from matplotlib import pyplot as pl

# Create 3 DF for three flowers

ds1 = ds[ds.target==0]

ds2 = ds[ds.target==1]

ds3 = ds[ds.target==2]

pl.scatter(ds1[‘sepal length (cm)’], ds1[‘sepal width (cm)’], color = ‘red’, marker = ‘o’)

pl.scatter(ds2[‘sepal length (cm)’], ds2[‘sepal width (cm)’], color = ‘blue’, marker = ‘o’)

pl.xlabel(‘Sepal Length’)

pl.ylabel(‘Sepal Width’)

pl.scatter(ds1[‘petal length (cm)’], ds1[‘petal width (cm)’], color = ‘red’, marker = ‘o’)

pl.scatter(ds2[‘petal length (cm)’], ds2[‘petal width (cm)’], color = ‘blue’, marker = ‘o’)

pl.xlabel(‘Petal Length’)

pl.ylabel(‘Petal Width’)

from sklearn.model_selection import train_test_split

# Remove fname column

X = ds.drop([‘target’,’fname’], axis = ‘columns’)

X.head()

y = ds[‘target’]

y.head()

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, shuffle=False)

from sklearn.svm import SVC

svmodel = SVC(kernel=‘linear’)

svmodel.fit(X_train, y_train)

SVC(kernel=‘linear’)

# Accuracy of model

svmodel.score(X_train, y_train)

0.9916666666666667