Treinando a arvore

train.py

# %%
import pandas as pd

df = pd.read_csv("../data/dados_pontos.csv",
                 sep=";")

# %%

from sklearn import model_selection
features = df.columns[3:-1]
target = 'flActive'

X_train, X_test, y_train, y_test = model_selection.train_test_split(df[features], 
                                                                    df[target],
                                                                    test_size=0.2,
                                                                    random_state=42,
                                                                    stratify=df[target]
                                                                    )

print("Tx Resposta Treino:", y_train.mean())
print("Tx Resposta Teste:", y_test.mean())


# %%
from sklearn import tree, metrics
# Arqui a gente treina
arvore = tree.DecisionTreeClassifier(max_depth=10, 
                                     min_samples_leaf=50,
                                     random_state=42)
arvore.fit(X_train, y_train)

# Aqui a gente prevê na propria base de treino
tree_pred_train = arvore.predict(X_train)
tree_acc_train = metrics.accuracy_score(y_train, tree_pred_train)
print("Arvore Train Acc :", tree_acc_train)

# Aqui a gente preve a base de teste
tree_pred_test = arvore.predict(X_test)
tree_acc_test = metrics.accuracy_score(y_test, tree_pred_test) #metrics.accuracy_score(y_test, tree_pred_train)
print("Arvore Test Acc:", tree_acc_test)

# probas
tree_proba_train = arvore.predict_proba(X_train)[:,1]
tree_acc_train_roc = metrics.roc_auc_score(y_train, tree_proba_train)
print("Arvore Train AUC:", tree_acc_train_roc)

tree_proba_test = arvore.predict_proba(X_test)[:,1]
tree_acc_test_roc = metrics.roc_auc_score(y_test, tree_proba_test)
print("Arvore Test AUC:", tree_acc_test_roc)

# %%
# Exemplo teo
y_test.mean()

y_teo = [0 for i in y_test]
acc_teo = metrics.accuracy_score(y_test, y_teo)
acc_teo
# %%