Classification(iris data with SGD and Hyper-parameter tuning)

[Notice] [ML_5]

Classification(iris data with SGD and Hyper-parameter tuning)

import warnings

# Avoid unnecessary warning output.
warnings.filterwarnings('ignore')

import pandas as pd
import seaborn as sns

from sklearn.datasets import load_iris

iris = load_iris()

data = iris['data']
feature_names = iris['feature_names']
target = iris['target']
iris['target_names']

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

Make DataFrame

df_iris = pd.DataFrame(data, columns=feature_names)

df_iris.head()

	sepal length (cm)	sepal width (cm)	petal length (cm)	petal width (cm)
0	5.1	3.5	1.4	0.2
1	4.9	3.0	1.4	0.2
2	4.7	3.2	1.3	0.2
3	4.6	3.1	1.5	0.2
4	5.0	3.6	1.4	0.2

df_iris['target'] = target

df_iris.head()

	sepal length (cm)	sepal width (cm)	petal length (cm)	petal width (cm)	target
0	5.1	3.5	1.4	0.2	0
1	4.9	3.0	1.4	0.2	0
2	4.7	3.2	1.3	0.2	0
3	4.6	3.1	1.5	0.2	0
4	5.0	3.6	1.4	0.2	0

from sklearn.model_selection import train_test_split

x_train, x_valid, y_train, y_valid = train_test_split(df_iris.drop('target', 1), df_iris['target'], stratify = df_iris['target'])

sns.countplot(y_train)

<AxesSubplot:xlabel='target', ylabel='count'>

x_train.shape, y_train.shape

((112, 4), (112,))

x_valid.shape, y_valid.shape

((38, 4), (38,))

SGDClassifier

from IPython.display import Image

# source: https://machinelearningnotepad.wordpress.com/
Image('https://machinelearningnotepad.files.wordpress.com/2018/04/yk1mk.png', width=500)

sklearn document

from sklearn.linear_model import SGDClassifier

sgd = SGDClassifier()

sgd.fit(x_train, y_train)

SGDClassifier()

prediction = sgd.predict(x_valid)

(prediction == y_valid).mean()

0.7894736842105263

Hyper-parameter tuning

document

• random_state: Fixed when hyperparameter tuning

• n_jobs=-1: use all CPU (fast learning speed)

sgd = SGDClassifier(penalty = 'elasticnet', random_state = 0, n_jobs = -1)

sgd.fit(x_train, y_train)

SGDClassifier(n_jobs=-1, penalty='elasticnet', random_state=0)

prediction = sgd.predict(x_valid)

(prediction == y_valid).mean()

0.6842105263157895