---
title: MLP Accessibility Score Predictor
emoji: 🚀
colorFrom: red
colorTo: pink
sdk: fastapi
app_file: app.py
requirements_file: requirements.txt
---

# MLP Accessibility Score Predictor

This Hugging Face Space hosts a FastAPI application that uses a pre-trained Multi-layer Perceptron (MLP) Regressor to predict urban accessibility scores. The model and its associated imputer are loaded directly from the Hugging Face Hub.

### How to use:

Send a POST request to the `/predict` endpoint with a JSON body containing the input features for which you want to predict the accessibility score. The expected features are based on the training data and include indicators such as `% ASF (Euclidean)`, `% Built-Up Area`, etc.

**Example Request Body (JSON):**
```json
{
  "perc_ASF_Euclidean": 0.6,
  "perc_Built_up_Area": 0.7,
  "perc_ASF_Network": 0.55,
  "perc_ASFS_from_Buffer_Distance_of_BS": 0.65,
  "Overall_Accessibility_Score": 0.62
}
```
(Note: The 'Overall_Accessibility_Score' is the target, but for prediction, you would typically provide values for the features and the model predicts the score. The `app.py` expects individual feature values.)

Access the API endpoint at `https://[your-space-name].hf.space/predict`.