Table 7.

Model parameters and hyperparameters, search space and optimization results for the convLSTM model.

Hyperparameter	Sampling	Search Space	Result
filter	categorical	$2^N,N\in \{4,\dots ,8\}$	128
kernelsize	categorical	$\in \{2,3,4\}$	3
dropout	log uniform distribution	$\in [0.01,1)$	0.22620
num segments	categorical	$\in \{2,4,5,8,10\}$	8
learning rate	log uniform distribution	$\in [0.001,0.1)$	0.09403