DecitionTreeのパラメータを調整する。
まずはMaxDepthから
from sklearn.model_selection import LeaveOneOut
from sklearn.metrics import accuracy_score
MAX_DEPTH = 20
depths = range(1, MAX_DEPTH)
loo_Y = d_train["Survived"].values
loo_X = d_train[["Pclass", "Sex", "Age", "Fare", "Parch", "Embarked", "SibSp"]].values
accuracy_scores = []
for depth in depths:
predicted_labels = []
loo = LeaveOneOut()
for train_index, test_index in loo.split(loo_X):
X_train, X_test = loo_X[train_index], loo_X[test_index]
y_train, y_test = loo_y[train_index], loo_y[test_index]
clf = DecisionTreeClassifier(max_depth=depth)
clf.fit(X_train, y_train)
predicted_label = clf.predict(loo_X[test_index])
predicted_labels.append(predicted_label)
score = accuracy_score(loo_Y, predicted_labels)
print('max depth={0}: {1}'.format(depth, score))
max depth=1: 0.7867564534231201
max depth=2: 0.6936026936026936
max depth=3: 0.8181818181818182
max depth=4: 0.8237934904601572
max depth=5: 0.8181818181818182
max depth=6: 0.8103254769921436
max depth=7: 0.8215488215488216
max depth=8: 0.8249158249158249
max depth=9: 0.8204264870931538
max depth=10: 0.8148148148148148
max depth=11: 0.8058361391694725
max depth=12: 0.8002244668911336
max depth=13: 0.797979797979798
max depth=14: 0.7934904601571269
max depth=15: 0.7912457912457912
max depth=16: 0.7755331088664422
max depth=17: 0.77665544332211
max depth=18: 0.7833894500561167
max depth=19: 0.7744107744107744
MaxDepthは8を利用する。
