2.3.2 Deep Learning Model
Deep learning model was constructed to improve the accuracy of the
predicted results31. In order to improve the accuracy
of the prediction results, a deep learning method is used to construct a
composite activity prediction model. The mean square error (MSE) is used
as a loss function. The neural network weights are updated using an Adam
optimizer with a learning rate of 0.006. Apply other parameters of the
original paper to the model. A four-layer fully connected neural network
is constructed by correcting the linear unit function as the activation
function. The Adam optimizer, whose learning rate 0.0001 and other
parameters are determined in the original file was used in this paper.
To reduce overfitting, the Dropout technique was applied to the second
layer (rate 0.4) and the third layer (rate 0.6) of the neural network.