The description of high-energy electromagnetic cascades in a highly granular digital calorimeter
Summary
We use data obtained with the Forward Calorimeter (FoCal) prototype to take the first steps in performing a three-dimensional fit on shower profiles. Within these shower profiles we observe saturation effects for the innermost rings around the central shower axis. To study the effect of saturation more thoroughly, we construct a toy model that simulates a single pixel sensor in the FoCal. Using a uniform distribution of incoming particles and a fixed cluster size of 3 on a grid of size 100x100, we observe a 2% deviation from linear behaviour at 100 simulated particles. The effect increases gradually towards a higher number of simulated particles. Finally, we use the toy model to generate no-photon and 1-photon data classes that we use for the Applied Data Science project. Using a Convolutional Neural Network we obtain a binary classification accuracy of 0.912 ± 0.063 after 10 epochs.