Task_3.py

import pandas as pd
import matplotlib.pyplot as plt
from sklearn.model_selection import train_test_split
from sklearn.compose import ColumnTransformer
from sklearn.preprocessing import OneHotEncoder
from sklearn.impute import SimpleImputer
from sklearn.pipeline import Pipeline
from sklearn.tree import DecisionTreeClassifier, plot_tree
from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score, confusion_matrix, classification_report
import joblib
import numpy as np

# 1) LOAD
# bank.csv from UCI is semicolon-delimited
df = pd.read_csv("bank.csv", sep=';')

# 2) TARGET & BASIC CLEANING
# Convert target y: 'yes'->1, 'no'->0
df['y'] = (df['y'] == 'yes').astype(int)

# IMPORTANT: 'duration' is known only after the call is finished -> data leakage.
# Drop it for modeling.
if 'duration' in df.columns:
    df = df.drop(columns=['duration'])

# 3) FEATURES / TYPES
y = df['y']
X = df.drop(columns=['y'])

numeric_features = X.select_dtypes(include=['int64', 'float64']).columns.tolist()
categorical_features = X.select_dtypes(include=['object']).columns.tolist()

# 4) PREPROCESSING
numeric_pipe = Pipeline([
    ("imputer", SimpleImputer(strategy="median"))
])

categorical_pipe = Pipeline([
    ("imputer", SimpleImputer(strategy="most_frequent")),   # handles 'unknown' too
    ("onehot", OneHotEncoder(handle_unknown="ignore"))
])

preprocess = ColumnTransformer([
    ("num", numeric_pipe, numeric_features),
    ("cat", categorical_pipe, categorical_features)
])

# 5) SPLIT
X_train, X_test, y_train, y_test = train_test_split(
    X, y, test_size=0.2, stratify=y, random_state=42
)

# 6) MODEL
clf = Pipeline([
    ("prep", preprocess),
    ("model", DecisionTreeClassifier(
        criterion="gini",
        max_depth=8,               # sensible defaults; keeps tree readable
        min_samples_split=20,
        min_samples_leaf=10,
        random_state=42
    ))
])

# 7) TRAIN
clf.fit(X_train, y_train)

# 8) EVALUATE
y_pred = clf.predict(X_test)

acc = accuracy_score(y_test, y_pred)
prec = precision_score(y_test, y_pred, zero_division=0)
rec = recall_score(y_test, y_pred, zero_division=0)
f1 = f1_score(y_test, y_pred, zero_division=0)

print("\n=== Metrics on Test Set ===")
print(f"Accuracy : {acc:.3f}")
print(f"Precision: {prec:.3f}")
print(f"Recall   : {rec:.3f}")
print(f"F1-score : {f1:.3f}\n")
print(classification_report(y_test, y_pred, target_names=["No", "Yes"]))

# 9) CONFUSION MATRIX (saved)
cm = confusion_matrix(y_test, y_pred)
fig = plt.figure(figsize=(5,4))
plt.imshow(cm, interpolation='nearest')
plt.title("Confusion Matrix")
plt.xticks([0,1], ["No","Yes"])
plt.yticks([0,1], ["No","Yes"])
for (i, j), v in np.ndenumerate(cm):
    plt.text(j, i, str(v), ha='center', va='center', fontsize=12)
plt.xlabel("Predicted")
plt.ylabel("Actual")
plt.tight_layout()
plt.savefig("confusion_matrix.png", dpi=200)
plt.close(fig)

# 10) OPTIONAL: TREE PLOT (feature names after one-hot)
# Get feature names from the preprocessor
oh = clf.named_steps["prep"].named_transformers_["cat"]["onehot"]
cat_names = oh.get_feature_names_out(categorical_features)
feat_names = np.r_[numeric_features, cat_names]

# Access the trained DecisionTree
tree_est = clf.named_steps["model"]

fig2 = plt.figure(figsize=(14, 8))
plot_tree(tree_est, feature_names=feat_names, class_names=["No","Yes"],
          filled=True, max_depth=3, fontsize=8)  # show top 3 levels for readability
plt.title("Decision Tree (Top Levels)")
plt.tight_layout()
plt.savefig("decision_tree.png", dpi=200)
plt.close(fig2)

# 11) SAVE MODEL
joblib.dump(clf, "dt_bank.pkl")
print("Saved: confusion_matrix.png, decision_tree.png, dt_bank.pkl")