# Osteoporosis Prediction Project

This project aims to develop a prediction system for osteoporosis using Logistic Regression and Random Forest models. The system predicts osteoporosis risk based on a variety of clinical and lifestyle factors. Below is a detailed explanation of the key components and scripts in this project.

## Documents Overview

**Treball_Final_Bioinfo.R**
This script contains the entire process of training and evaluating a Logistic Regression model for osteoporosis prediction. The key steps include:

1. Package Installation
   
2. Dataset Preparation:
   - Loads the `osteoporosis.csv` dataset, removes irrelevant columns, handles missing data, and converts categorical variables into factors.
   
3. Model Training:
   - Trains a Logistic Regression model using 70% of the dataset (training set) and evaluates it using confusion matrix, ROC curve, and Precision-Recall curve.

4. Feature Importance Evaluation:
   - A Random Forest model is trained to identify the most influential features for predicting osteoporosis.

5. Model Saving:
   - Saves the trained Logistic Regression and Random Forest models for future use.

**API.R**
This script sets up a **REST API** using the plumber package to provide osteoporosis predictions based on the trained Logistic Regression model. The key steps are:

1. Library Loading

2. API Endpoint Definition:
   - Defines the predict_logistic endpoint that accepts user inputs such as gender, age, family history, physical activity, etc., and returns the probability and prediction (Yes/No) of osteoporosis.

3. Input Format:
   - Accepts both categorical (e.g., gender, family history) and numeric (e.g., age) input variables for making predictions.

4. Prediction:
   - Uses the trained Logistic Regression model to make predictions based on the provided inputs.

**Run_API.R**
This script runs the API defined in 'API.R' on a local server. 

---

## How to Run the Project

1. **Train the model**:
   - Run 'Treball_Final_Bioinfo.R' to train the Logistic Regression. Make sure the dataset (osteoporosis.csv) is placed correctly at the specified path.

2. **Run the API**:
   - Run 'API.R' to load the trained Logistic Regression model and set up the API with the predict_logistic` endpoint.

3. **Start the server**:
   - Run 'Run_API.R' to start the server and make the API available on port 8000.

---

## Required Packages

The following R packages are required to run the project:

- `plumber`: For creating REST APIs in R.
- `caret`: For building predictive models.
- `dplyr`: For data manipulation.
- `PRROC`: For Precision-Recall curve creation.
- `pROC`: For ROC curve creation and evaluation.
- `smotefamily`: For handling imbalanced data through SMOTE.
- `randomForest`: For building Random Forest models.