Module 2: Apprentissage supervisé

• Partie 2.1: Les K plus Proches Voisins [Notebook]
• Partie 2.2: Régression linéaire [Notebook]
• Partie 2.3: Régression logistique [Notebook]
• Partie 2.4: SVM [Notebook]

Partie 2.1: KNN - Les K plus Proches Voisins

Le knn est un algorithme d'apprentissage supervisé

import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
from sklearn.neighbors import KNeighborsRegressor, KNeighborsClassifier

iris = pd.read_csv("data/iris.csv")
iris.head()

	sepal_length	sepal_width	petal_length	petal_width	species
0	5.1	3.5	1.4	0.2	setosa
1	4.9	3.0	1.4	0.2	setosa
2	4.7	3.2	1.3	0.2	setosa
3	4.6	3.1	1.5	0.2	setosa
4	5.0	3.6	1.4	0.2	setosa

plt.figure(figsize=(10, 8))
sns.scatterplot(x="petal_width", y="petal_length", hue="species", data=iris)
plt.show()

Classification

Dans le cadre d'une classification, on va classifier les points en fonction de leur espèce.

X = dt.petal_width.values.reshape(-1, 1)
y = dt.petal_length.values

from sklearn.model_selection import train_test_split
X_train, X_test, y_train, y_test = train_test_split(X, y, shuffle=True, test_size=.3)

knn = KNeighborsRegressor(n_neighbors=5)

knn.fit(X_train, y_train)

KNeighborsRegressor(algorithm='auto', leaf_size=30, metric='minkowski',
                    metric_params=None, n_jobs=None, n_neighbors=5, p=2,
                    weights='uniform')

Évaluation

from sklearn.metrics import r2_score
y_pred = knn.predict(X_test)
score = r2_score(y_test, y_pred)
score

0.9377572372679783

plt.scatter(y_test, y_pred)
plt.plot(y_test, y_test, c="red")

[<matplotlib.lines.Line2D at 0x7ff07e363090>]

k_values = [2, 3, 4, 5, 6, 7, 8]

train_scores = []
test_scores = []

for k in k_values:
    knn = KNeighborsRegressor(n_neighbors=k)
    knn.fit(X_train, y_train)
    train_scores.append(knn.score(X_train, y_train))
    test_scores.append(knn.score(X_test, y_test))

plt.plot(k_values, train_scores)
plt.plot(k_values, test_scores)

[<matplotlib.lines.Line2D at 0x7ff07e337e50>]

X = dt[["petal_width", "petal_length"]]
species = dt["species"]

from sklearn.preprocessing import LabelEncoder
encoder = LabelEncoder()
y = encoder.fit_transform(species)

from sklearn.model_selection import train_test_split
X_train, X_test, y_train, y_test = train_test_split(X, y, shuffle=True, test_size=.3)

clf = KNeighborsClassifier(n_neighbors=5)

clf.fit(X_train, y_train)

KNeighborsClassifier(algorithm='auto', leaf_size=30, metric='minkowski',
                     metric_params=None, n_jobs=None, n_neighbors=5, p=2,
                     weights='uniform')

Évaluation

from sklearn.metrics import accuracy_score
y_pred = clf.predict(X_test)
score = accuracy_score(y_test, y_pred)
score

0.9333333333333333

k_values = [2, 3, 4, 5, 6, 7, 8]

train_scores = []
test_scores = []

for k in k_values:
    knn = KNeighborsClassifier(n_neighbors=k)
    knn.fit(X_train, y_train)
    train_scores.append(knn.score(X_train, y_train))
    test_scores.append(knn.score(X_test, y_test))

plt.plot(k_values, train_scores)
plt.plot(k_values, test_scores)

[<matplotlib.lines.Line2D at 0x7ff07e2beb90>]

Classification

Dans le cadre d'une classification

clf.predict([[1.5, 5.0]]), clf.predict_proba([[1.5, 5.0]])

(array(['versicolor'], dtype=object), array([[0. , 0.8, 0.2]]))

sns.pairplot(dt, hue='species')

<seaborn.axisgrid.PairGrid at 0x7ff07d83c610>