Authors: [V. Belavin] (vbelavin@hse.ru), [E. Trofimova] (etrofimova@hse.ru), [A. Ustyuzhanin] (austyuzhanin@hse.ru)
This directory contains code necessary to run the Electromagnetic Showers (EM) Reconstruction algorithm that is devided into the following parts:
- Graph Construction;
- Edge Classification;
- Showers Clusterization;
- Parameters Reconstruction.
X, Y, Z coordinates and the direction of the EM Showers base-tracks.
The showers are generated using FairShip framework.
Data for graph generation is located here: https://gitlab.com/SchattenGenie/shower_generation/blob/master/data/mcdata_taue2.root
The algorithm detects ~ 86% of Showers and assess the coordinates and direction of base-tracks with ~ 75% accuracy.
training_classifier.py predicts the probability of edge to connect vertices of one shower. The probability will be then used as edge weight for proposed clusterization algorithm in clustering.py.