# -*- coding: utf-8 -*-
"""project ml.ipynb

Automatically generated by Colab.

Original file is located at
    https://colab.research.google.com/drive/1ScYdmUla5mT7Zrevk6Lk9LVw2kmRfChM
"""

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns

from sklearn.model_selection import train_test_split
from sklearn.linear_model import LogisticRegression
from sklearn.ensemble import RandomForestClassifier
from sklearn.metrics import accuracy_score, classification_report, confusion_matrix
from imblearn.over_sampling import SMOTE
from xgboost import XGBClassifier
from sklearn.neural_network import MLPClassifier
from sklearn.svm import SVC
from sklearn.neighbors import KNeighborsClassifier

df=pd.read_csv("healthcare_stroke_dataset.csv")
df.head()

df.info()

df.shape

df.describe()

df.isnull().sum()

df['bmi'].fillna(df['bmi'].mean(), inplace=True)
df.isnull().sum()

df.describe()

plt.figure(figsize=(12,8))
a = ['age', 'avg_glucose_level', 'bmi']
for i, c in enumerate(a):
    plt.subplot(2, 2, i+1)
    sns.histplot(df[c], kde=True, bins=30)
    plt.title(f'Distribution of {c}')

plt.tight_layout()
plt.show()

plt.figure(figsize=(12,6))

s = ['age', 'avg_glucose_level', 'bmi']
for i, o in enumerate(s):
    plt.subplot(1, 3, i+1)
    sns.boxplot(y=df[o])
    plt.title(f'Boxplot of {o}')

plt.tight_layout()
plt.show()

if 'id' in df.columns:
    df = df.drop('id', axis=1)

nu= df.select_dtypes(include=np.number)

corr = nu.corr()

plt.figure(figsize=(10,6))
sns.heatmap(corr, annot=True, cmap='coolwarm', fmt=".2f")
plt.title('Correlation Heatmap')
plt.show()

dfe = pd.get_dummies(df, drop_first=True)

dfe.head()

df = df.drop('date', axis=1)

df_encoded = pd.get_dummies(df, drop_first=True)

df_encoded.head()


y = df_encoded['stroke']

X = df_encoded.drop('stroke', axis=1)

# تقسيم البيانات: 80% تدريب و20% اختبار
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)



print("Configure training data:", X_train.shape)
print("Test data format:", X_test.shape)

smote = SMOTE(random_state=42)

X_train_res, y_train_res = smote.fit_resample(X_train, y_train)

print("📊 Before budget:")
print(y_train.value_counts())
print("\n📊 After budget:")
print(y_train_res.value_counts())

model_lr  = LogisticRegression(class_weight='balanced', max_iter=3000)
model_rf  = RandomForestClassifier(class_weight='balanced', random_state=42)
model_xgb = XGBClassifier(scale_pos_weight=(len(y_train_res) - sum(y_train_res)) / sum(y_train_res))
model_knn = KNeighborsClassifier(n_neighbors=5)
model_mlp = MLPClassifier(hidden_layer_sizes=(100,), max_iter=500, random_state=42)
model_svc = SVC(class_weight='balanced', probability=True, random_state=42)

from sklearn.model_selection import cross_val_score

models = {
    "Logistic Regression": model_lr,
    "Random Forest": model_rf,
    "XGBoost": model_xgb,
    "KNN": model_knn,
    "MLP": model_mlp,
    "SVC": model_svc
}

results = {}

cv_folds = 5

for name, model in models.items():
    print("="*60)
    print(f"🤖 Training Model with Cross Validation: {name}")
    print("="*60)

    cv_scores = cross_val_score(model, X_train_res, y_train_res, cv=cv_folds, scoring='f1')
    print(f"Cross Validation F1 Scores: {cv_scores}")
    print(f"Average CV F1 Score: {cv_scores.mean():.3f}")

    model.fit(X_train_res, y_train_res)

    y_pred = model.predict(X_test)

    acc = accuracy_score(y_test, y_pred)
    report = classification_report(y_test, y_pred, digits=3, output_dict=True)
    conf = confusion_matrix(y_test, y_pred)

    results[name] = {
        "Accuracy": acc,
        "Precision_1": report["1"]["precision"],
        "Recall_1": report["1"]["recall"],
        "F1_1": report["1"]["f1-score"],
        "CV_F1_Avg": cv_scores.mean()
    }

    print(f"Accuracy: {acc:.3f}")
    print(f"Precision (Class 1): {report['1']['precision']:.3f}")
    print(f"Recall (Class 1): {report['1']['recall']:.3f}")
    print(f"F1-score (Class 1): {report['1']['f1-score']:.3f}")
    print(f"Confusion Matrix:\n{conf}")
    print("-"*60, "\n")

results_df = pd.DataFrame(results).T
print("\n📊 Model Comparison (After Cross Validation):")
display(results_df)