Update app.py

app.py

@@ -5,7 +5,6 @@ from sklearn.preprocessing import LabelEncoder, MinMaxScaler
 from sklearn.model_selection import train_test_split
 from keras.models import Sequential
 from keras.layers import Dense, Dropout
-import gradio as gr
 
 data = pd.read_csv('cars_raw.csv')
 
@@ -29,6 +28,7 @@ for col in data.select_dtypes(include=['category', 'object']).columns:
 for col in data.select_dtypes(include=['number']).columns:
     data[col] = scaler.fit_transform(data[col].values.reshape(-1, 1))
 
+# Удаление ненужных колонок
 data = data.drop(columns=["Mileage", "SellerType", "VIN", "Stock#", "Drivetrain", "SellerName", "ConsumerReviews", "ExteriorStylingRating", "State", "Zipcode", "DealType"])
 data_df = pd.DataFrame(data)
 
@@ -46,9 +46,12 @@ model.add(Dense(1))
 model.compile(optimizer='adam', loss='mean_squared_error')
 model.fit(X_train, y_train, epochs=100, batch_size=32, validation_data=(X_test, y_test))
 
+# Оценка модели
 test_loss = model.evaluate(X_test, y_test)
 print(f'Test loss (MSE): {test_loss}')
 
+import gradio as gr
+
 def predict_car_price(Year, Make, Model, FuelType, Engine):
     car_data = {
         "Year": int(Year),
@@ -59,21 +62,30 @@ def predict_car_price(Year, Make, Model, FuelType, Engine):
     }
 
     car_data_df = pd.DataFrame([car_data])
-    needs_df = data.drop(columns=["Price"]).copy()
+    result_data = pd.read_csv('cars_raw.csv')
+    result_data_df = pd.DataFrame(result_data)
+
+    needs_df = result_data_df.drop(columns=["Mileage", "SellerType", "VIN", "Stock#", "Drivetrain", "SellerName", "ConsumerReviews", "ExteriorStylingRating", "State", "Zipcode", "DealType", "Used/New", "Price", "ConsumerRating", "SellerRating", "SellerReviews", "StreetName", "ComfortRating", "PerformanceRating", "ValueForMoneyRating", "ReliabilityRating", "ExteriorColor", "InteriorColor", "MinMPG", "MaxMPG", "Transmission", "InteriorDesignRating"])
     needs_df = pd.concat([needs_df, car_data_df], ignore_index=True)
 
+    le = LabelEncoder()
     for col in needs_df.select_dtypes(include=['category', 'object']).columns:
         needs_df[col] = le.fit_transform(needs_df[col])
 
+    scaler = MinMaxScaler()
     for col in needs_df.select_dtypes(include=['number']).columns:
-        needs_df[col] = scaler.transform(needs_df[col].values.reshape(-1, 1))
+        needs_df[col] = scaler.fit_transform(needs_df[col].values.reshape(-1, 1))
+
+    result_df = needs_df.loc[[0]]
 
-    result_df = needs_df.iloc[-1].values.reshape(1, -1)
     prediction = model.predict(result_df)
-    inverted_prediction = scaler.inverse_transform(prediction.reshape(-1, 1))
+
+    inverted_prediction = scaler.inverse_transform(
+        prediction.reshape(-1, 1)
+    )
     predicted_price = inverted_prediction[0][0]
 
-    return round(predicted_price, 2)
+    return predicted_price
 
 iface = gr.Interface(
     fn=predict_car_price,