Gamma-Ray Burst Detection from Radio Signals Using CNN-LSTM

This repository contains the code for a project on detecting gamma-ray bursts (GRBs) using neural networks (CNN + LSTM), processing radio signal data.

Project Structure

• Data/
  Contains:

  • grb_txt/: raw signal data (txt) from the original dataset.
  • GRL2015.pdf: the scientific paper we relied on; the source of the dataset.

• data_preprocesing/
  Code for data preparation, to be run in the following order:

  1. txt_to_csv.ipynb: converts raw txt files to csv format.
  2. detect_peak.ipynb: detects peaks and creates a readable table.
  3. data_for_uniq.ipynb or data_for_3sec.ipynb: prepares datasets for 30-second or 5-second bins.
  4. united_sample.ipynb or united_3sec.ipynb: merges all data into a final table.
  5. grafs.ipynb: visualizes graphs for comparison with results from GRL2015.pdf.

• model_training/
  Contains:

  • experiment_1.ipynb, experiment2.ipynb, experiment3.ipynb: code for three experiments described in the paper.
  • experiment_1.h5, experiment_2.h5, experiment_3.h5: saved trained models for each experiment.

Usage

The workflow is designed to be run in Jupyter Notebook. Simply follow the notebook sequence as described.