LSTM_classification

Classification of Multi-particle events using LSTM's and Convolutional Neural Networks.

Data

Generated in Pythia8240
Data is stored in .txt files, where each line represents an event. Each event is comprised of a list of particle variables (pT, eta, phi), which are separated by commas.
The data is split into two folders, ATLAS-like (LHC) and STAR-like (RHIC). Each of those datasets coontaint MinBias (ND), Difr (SD+DD), SinDifr (SD) and DblDifr (DD).

Neural Nets

2class_conventional.ipynb == rapidity gap dense net for bg/sig classification
2class_hist_CNN.ipynb == pT histogram convnet for bg/sig classification
2class_LSTM_var.ipynb == var length LSTM net for bg/sig classification
3class_conventional.ipynb == rapidity gap dense net for ND/SD/DD classification
3class_LSTM_var.ipynb == var length LSTM net for ND/SD/DD classification
2class_LSTM_padded.py = fixed length LSTM net for bg/sig classification
3class_LSTM_padded.py = fixed length LSTM net for ND/SD/DD classification
2class_cnn.py = Binary mapped histogram convnet for bg/sig classification
3class_cnn.py = Binary mapped histogram convnet for ND/SD/DD classification\