Çapraz Doğrulama Teknikleri – Cross Validation

Merhaba iyi günler.

Bugün; popülaritesi gün geçtikçe yükselen, birçok makine öğrenmesi-yapay zeka temelli uygulamalarda açık ara ön planda olan, veri madenciliği-bilimsel analiz çözümlemelerinde esnek çözümlemeler sunan, hızlı ve çevik yapısı ile kullanıcıların gözdesi olarak yer edinmiş olan Python dili üzerinde bir modelin performansının nasıl olduğunun tahmin edilmesi için kullanılan Çapraz Doğrulama – Cross Validation çözümlerini inceliyor olacağım.

Şimdiden iyi okumalar.

İlgili yazı içeriğine ulaşmak için lütfen tıklayınız...

Please click to access the related article content...

İlgili derlenmiş kod çözümü ise;

The relevant compiled code solution is;

##################################
#Randomized/Monte Carlo - MCCV Cross Validation - Rastgele Capraz Dogrulama
from sklearn.model_selection import ShuffleSplit

X = [...] # Veri kumesi verileri. / Ornegin; Iris=Setosa=0.42 ... vb.
y = [...] # Veri kumesi etiketleri. / Ornegin; Iris=Setosa,Versicolor... vb.

rs = ShuffleSplit(n_splits=5, test_size=0.2, random_state=0)

for train_index, test_index in rs.split(X):
    X_train, X_test = X[train_index], X[test_index]
    y_train, y_test = y[train_index], y[test_index]
    model = ...  # Model
    model.fit(X_train, y_train)
    accuracy = model.score(X_test, y_test)
    print(accuracy)

##################################
#K-Fold Cross Validation - K Katli/Katmanli Capraz Dogrulama
from sklearn.model_selection import KFold

X = [...] # Veri kumesi verileri. / Ornegin; Iris=Setosa=0.42 ... vb.
y = [...] # Veri kumesi etiketleri. / Ornegin; Iris=Setosa,Versicolor... vb.

kf = KFold(n_splits=5)

for train_index, test_index in kf.split(X):
    X_train, X_test = X[test_index]
    y_train, y_test = y[train_index], y[test_index]
    model = ...  # Model
    model.fit(X_train, y_train)
    accuracy = model.score(X_test, y_test)
    print(accuracy)

##################################
#Leave One/P Out Cross Validation - Biri (P) Disinda/Disarida Bırakilan Capraz Dogrulama
from sklearn.model_selection import LeaveOneOut

X = [...] # Veri kumesi verileri. / Ornegin; Iris=Setosa=0.42 ... vb.
y = [...] # Veri kumesi etiketleri. / Ornegin; Iris=Setosa,Versicolor... vb.

loo = LeaveOneOut()

for train_index, test_index in loo.split(X):
    X_train, X_test = X[train_index], X[test_index]
    y_train, y_test = y[train_index], y[test_index]
    model = ...  # Model
    model.fit(X_train, y_train)
    accuracy = model.score(X_test, y_test)
    print(accuracy)

##################################
#Simple Cross Validation - Basit Capraz Dogrulama
from sklearn.model_selection import train_test_split

X = [...] # Veri kumesi verileri. / Ornegin; Iris=Setosa=0.42 ... vb.
y = [...] # Veri kumesi etiketleri. / Ornegin; Iris=Setosa,Versicolor... vb.

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

model = ...  # Model
model.fit(X_train, y_train)
accuracy = model.score(X_test, y_test)
print(accuracy)

##################################
#Repeated K-Fold Cross Validation - Tekrarli K Katmanli Capraz Dogrulama
from sklearn.model_selection import RepeatedKFold

X = [...] # Veri kumesi verileri. / Ornegin; Iris=Setosa=0.42 ... vb.
y = [...] # Veri kumesi etiketleri. / Ornegin; Iris=Setosa,Versicolor... vb.

rkf = RepeatedKFold(n_splits=5, n_repeats=10)

for train_index, test_index in rkf.split(X):
    X_train, X_test = X[train_index], X[test_index]
    y_train, y_test = y[train_index], y[test_index]
    model = ...  # Model
    model.fit(X_train, y_train)    
    accuracy = model.score(X_test, y_test)
    print(accuracy)

##################################
#Holdout Cross Validation - Ayirarak Capraz Dogrulama
from sklearn.model_selection import train_test_split

X = [...] # Veri kumesi verileri. / Ornegin; Iris=Setosa=0.42 ... vb.
y = [...] # Veri kumesi etiketleri. / Ornegin; Iris=Setosa,Versicolor... vb.

X_train, X_test, y_train, y_test = train_test_split(X, y ,test_size=0.2 ,random_state=0)

model = ...  # Model
model.fit(X_train, y_train)
accuracy = model.score(X_test, y_test)
print(accuracy)

##################################
#Rolling/Rolling Window Cross Validation - Kaydirmali/Kaydirmali Bakis ile Capraz Dogrulama
from sklearn.model_selection import TimeSeriesSplit

X = [...] # Veri kumesi verileri. / Ornegin; Iris=Setosa=0.42 ... vb.
y = [...] # Veri kumesi etiketleri. / Ornegin; Iris=Setosa,Versicolor... vb.

tscv = TimeSeriesSplit(n_splits=5)

for train_index, test_index in tscv.split(X):
    X_train, X_test = X[train_index], X[test_index]
    y_train, y_test = y[train_index], y[test_index]
    model = ...  # Model
    model.fit(X_train, y_train)
    accuracy = model.score(X_test, y_test)
    print(accuracy)

##################################
#Bootstrap Cross Validation - Onyuklemeli Capraz Dogrulama
from sklearn.utils import resample

X = [...] # Veri kumesi verileri. / Ornegin; Iris=Setosa=0.42 ... vb.
y = [...] # Veri kumesi etiketleri. / Ornegin; Iris=Setosa,Versicolor... vb.

for i in range(10):
    X_resample, y_resample = resample(X, y)
    model = ...  # Model
    model.fit(X_resample, y_resample)
    accuracy = model.score(X, y)
    print(accuracy)