objectif

Classification par plus proches voisins.

code minimal

from sklearn.datasets import load_iris
from sklearn.neighbors import KNeighborsClassifier
X, y = load_iris(return_X_y=True)
clf = KNeighborsClassifier(n_neighbors=3).fit(X, y)
print(len(set(clf.predict(X)[:5])) >= 1)

utilisation

from sklearn.neighbors import KNeighborsRegressor
print(hasattr(KNeighborsRegressor(n_neighbors=2), "fit"))

variante(s) utile(s)

from sklearn.neighbors import NearestNeighbors
nbrs = NearestNeighbors(n_neighbors=2).fit([[0,0],[1,1]])
print(nbrs.kneighbors([[0,0]])[0].shape[1])

notes

• Standardiser les features; choisir k impair pour binaire.