📊 Sims_calibrate

This repository contains R scripts for parameter generation, model loading, calibration, prediction, and analysis for epidemic simulation studies.
Whether you want to run simulations 🧮, calibrate predictions 🔍, or visualize epidemic curves 📈 — this project has you covered.

---

📂 File Descriptions

00-params.R

📝 Generates the simulation parameters used throughout the analysis.

01a-bilstm.R

🤖 Loads the pre-trained BiLSTM model for epidemic curve prediction.

01b-abc.R

🎯 Applies Approximate Bayesian Computation (ABC) for calibration.

02-abc-bilstm-prediction.R

🔗 Combines ABC calibration with BiLSTM predictions and saves the results.

03-epicurves-stats.R

📈 Generates epidemic curves from the predicted parameters.

04-parameter-stats.R

📊 Analyzes predicted parameters to find bias and create boxplots.

---

📦 Additional Resources

LSTM_model/

📁 Contains the pre-trained BiLSTM model used for machine learning–based calibration.
These models are loaded by 01a-bilstm.R and used to improve prediction accuracy.

model_evaluation/

📁 Contains PyTorch code for reading and processing epidemic simulation results:

• Loads incidence.csv 📄 (epidemic incidence data)
• Loads theta.csv 📄 (model parameter values)
• Trains and evaluates a PyTorch-based deep learning model for epidemic forecasting.

---

🛠️ Project Notes

• R version: Ensure you’re using a compatible R version 📦 with all required packages installed.
• Execution order: Run scripts in the order listed above for consistent results.
• Outputs: Predictions, epidemic curves, and statistical plots are saved in the designated output folders.

---

🚀 Getting Started

1️⃣ Generate parameters

Rscript 00-params.R

2️⃣ Load model & run calibration

Rscript 01a-bilstm.R
Rscript 01b-abc.R

3️⃣ Combine calibration & predictions

Rscript 02-abc-bilstm-prediction.R

4️⃣ Generate epicurves

Rscript 03-epicurves-stats.R

5️⃣ Analyze parameter statistics

Rscript 04-parameter-stats.R