Create app.py

app.py

@@ -0,0 +1,99 @@
+import pandas as pd
+from sklearn.model_selection import train_test_split
+from sklearn.linear_model import LogisticRegression
+from sklearn.ensemble import RandomForestClassifier
+from sklearn.metrics import classification_report
+import gradio as gr
+
+# Sabit CSV dosya yolu
+CSV_FILE_PATH = "payment_fraud.csv"  # DOSYA ADI GÜNCELLENDİ
+
+# Veri setini yükleme ve ön işleme fonksiyonu
+def load_and_preprocess():
+    try:
+        data = pd.read_csv(CSV_FILE_PATH)
+        data["type"] = data["type"].map({"CASH_OUT": 1, "PAYMENT": 2, "CASH_IN": 3, "TRANSFER": 4, "DEBIT": 5})
+        data["isFraud"] = data["isFraud"].map({0: "No Fraud", 1: "Fraud"})
+
+        # Sahtekarlık oranını hesapla
+        fraud_rate = (data["isFraud"].value_counts(normalize=True) * 100).get("Fraud", 0)
+
+        return data, fraud_rate
+    except FileNotFoundError:
+        print(f"Hata: {CSV_FILE_PATH} dosyası bulunamadı!")
+        return None, 0  # Hata durumunda fraud_rate 0 olarak döner
+
+# Modeli eğitme ve değerlendirme fonksiyonu
+def train_and_evaluate(data):
+    if data is None:
+        return None, None, "Veri yükleme hatası nedeniyle model eğitilemedi."
+
+    x = data[["type", "amount", "oldbalanceOrg", "newbalanceOrig"]].values
+    y = data[["isFraud"]].values
+    xtrain, xtest, ytrain, ytest = train_test_split(x, y, test_size=0.10, random_state=42)
+
+    # Logistic Regression Modeli
+    model_lr = LogisticRegression()
+    model_lr.fit(xtrain, ytrain.ravel())
+    y_pred_lr = model_lr.predict(xtest)
+    report_lr = classification_report(ytest, y_pred_lr)
+
+    # Random Forest Modeli
+    model_rf = RandomForestClassifier()
+    model_rf.fit(xtrain, ytrain.ravel())
+    y_pred_rf = model_rf.predict(xtest)
+    report_rf = classification_report(ytest, y_pred_rf)
+
+    return model_lr, model_rf, report_lr, report_rf
+
+# Sahtekarlık tespiti ve olasılık hesabı
+def fraud_detection(models, type, amount, oldbalanceOrg, newbalanceOrig):
+    model_lr, model_rf = models
+    if model_lr is None or model_rf is None:
+        return "Model eğitilemediği için tahmin yapılamıyor."
+
+    features = [[type, amount, oldbalanceOrg, newbalanceOrig]]
+
+    # Logistic Regression tahmini
+    prediction_lr = model_lr.predict(features)[0]
+    probability_lr = model_lr.predict_proba(features)[0][1] if hasattr(model_lr, "predict_proba") else "N/A"
+
+    # Random Forest tahmini
+    prediction_rf = model_rf.predict(features)[0]
+    probability_rf = model_rf.predict_proba(features)[0][1] if hasattr(model_rf, "predict_proba") else "N/A"
+
+    return f"Logistic Regression: {prediction_lr} (Olasılık: {probability_lr})\nRandom Forest: {prediction_rf} (Olasılık: {probability_rf})"
+
+# Gradio Arayüzü
+def analyze_and_predict(type, amount, oldbalanceOrg, newbalanceOrig):
+    data, fraud_rate = load_and_preprocess()
+    model_lr, model_rf, report_lr, report_rf = train_and_evaluate(data)
+
+    # Analiz sonuçları
+    analysis_results = f"Logistic Regression Raporu:\n{report_lr}\n\nRandom Forest Raporu:\n{report_rf}\n\nSahtekarlık Oranı: {fraud_rate:.2f}%"
+
+    # Tahmin Sonucu
+    prediction_result = fraud_detection((model_lr, model_rf), type, amount, oldbalanceOrg, newbalanceOrig)
+
+    return analysis_results, prediction_result
+
+# Gradio Arayüzü
+with gr.Blocks() as demo:
+    with gr.Row():
+        with gr.Column():
+            type = gr.Dropdown(label="İşlem Türü", choices=[("CASH_OUT", 1), ("PAYMENT", 2), ("CASH_IN", 3), ("TRANSFER", 4), ("DEBIT", 5)], type="value")
+            amount = gr.Number(label="İşlem Miktarı")
+            oldbalanceOrg = gr.Number(label="Gönderen Eski Bakiye")
+            newbalanceOrig = gr.Number(label="Gönderen Yeni Bakiye")
+            predict_button = gr.Button("Analiz Et ve Tahmin Yap")
+        with gr.Column():
+            analysis_output = gr.Textbox(label="Analiz Sonuçları", lines=5)
+            prediction_output = gr.Textbox(label="Tahmin Sonucu")
+
+    predict_button.click(
+        fn=analyze_and_predict,
+        inputs=[type, amount, oldbalanceOrg, newbalanceOrig],
+        outputs=[analysis_output, prediction_output]
+    )
+
+demo.launch(debug=True, share=True)