Skip to main content

Table 2 (abstract P5). Best neural network parameterization for prosumers and non-prosumers

From: Abstracts from the 9th DACH+ Conference on Energy Informatics

 

Non-Prosumer Prosumer

Optimizer

Adam

Adam

Learning rate

0.001

0.001

Loss function

MAE

MSE

Number of hidden layers

3

5

Number of neurons per hidden layer

1000

500

Activation function

ReLu

ReLu

Early stopping algorithm patience

20

20

Validation split

0.1

0.1

Combination of lagged input variables

7

4