updated app and requirements

app.py

@@ -1,48 +1,62 @@
 import gradio as gr
+import pandas as pd
 import joblib
-import json
-import numpy as np
 
 # Load the trained model
-model = joblib.load('trained_model.pkl')
-
-# Load feature names
-with open('selected_features.json', 'r') as f:
-    feature_names = json.load(f)
-
-# Define prediction function
-def predict_heart_failure(*args):
-    """
-    Predict heart failure likelihood based on input features.
-
-    Args:
-        *args: Input values for each feature.
-
-    Returns:
-        str: Prediction result.
-        str: Prediction probability as a percentage.
-    """
-    input_data = np.array(args).reshape(1, -1)
-    prediction = model.predict(input_data)
-    probability = model.predict_proba(input_data)
-    result = "Likely" if prediction[0] == 1 else "Unlikely"
-    return result, f"{probability[0][1] * 100:.2f}%"
-
-# Create Gradio interface
-inputs = [gr.inputs.Number(label=feature) for feature in feature_names]
-outputs = [
-    gr.outputs.Textbox(label="Prediction"),
-    gr.outputs.Textbox(label="Probability (%)")
+model_path = "trained_model.pkl"
+rf_model = joblib.load(model_path)
+
+# Define feature ranges and labels based on data
+numerical_features = ['BMI', 'WeightInKilograms', 'HeightInMeters', 'PhysicalHealthDays', 'SleepHours']
+categorical_features = [
+    'HadAngina_Yes', 'HadHeartAttack_Yes', 'ChestScan_Yes', 
+    'HadStroke_Yes', 'DifficultyWalking_Yes', 'HadDiabetes_Yes', 
+    'PneumoVaxEver_Yes', 'HadArthritis_Yes'
 ]
 
-app = gr.Interface(
-    fn=predict_heart_failure,
+# Define sliders for numerical features
+sliders = {
+    "BMI": (0, 50, 1),
+    "WeightInKilograms": (30, 200, 1),
+    "HeightInMeters": (1.0, 2.5, 0.01),
+    "PhysicalHealthDays": (0, 30, 1),
+    "SleepHours": (0, 24, 1)
+}
+
+# Define radio buttons for categorical features
+radio_options = ['Yes', 'No']
+
+# Prediction function
+def predict_outcome(*inputs):
+    input_data = dict(zip(numerical_features + categorical_features, inputs))
+    
+    # Convert categorical inputs to numerical
+    for feature in categorical_features:
+        input_data[feature] = 1 if input_data[feature] == "Yes" else 0
+    
+    # Create input DataFrame
+    input_df = pd.DataFrame([input_data])
+    
+    # Predict using the model
+    prediction = rf_model.predict(input_df)[0]
+    return "High Risk" if prediction == 1 else "Low Risk"
+
+# Build Gradio interface
+inputs = [
+    gr.Slider(sliders[feature][0], sliders[feature][1], sliders[feature][2], label=feature) 
+    for feature in numerical_features
+] + [
+    gr.Radio(radio_options, label=feature) for feature in categorical_features
+]
+
+interface = gr.Interface(
+    fn=predict_outcome,
     inputs=inputs,
-    outputs=outputs,
-    title="Heart Failure Prediction",
-    description="Enter the values for the features to predict the likelihood of heart failure."
+    outputs="text",
+    title="Health Risk Prediction",
+    description="Predicts health risks based on input parameters using the trained model."
 )
 
 # Launch the app
 if __name__ == "__main__":
-    app.launch()
+    interface.launch()

requirements.txt

@@ -2,3 +2,4 @@ gradio
 numpy
 joblib
 scikit-learn
+pandas