Table 2. Variable Values before and after Optimization of the LSTM Network Layer Functions.

function	variable	value before optimization	optimized value	describe
sequenceInputLayer	InputSize	9	9	number of input parameters
	MinLength	1	1	minimum input length of data
lstmLayer	InputSize	auto	auto	automatic adjustment
	NumHiddenUnits	100, 50	128, 64	LSTM hidden layer
	Bias	32	32	offset term
dropoutLayer	dropout	0.2	0.2	probability of node dropping
fullyConnectedLayer	OutputSize	1	1	number of output parameters
regressionLayer				regression output
trainingOptions	MaxEpochs	100	50	number of training set runs
	InitialLearnRate	0.005	0.001	learning rate
	MinBatchSize	200	170	amount of data per training