Method, information processing device, and recording medium storing instructions for prediction
A method for prediction executed by an information processing device includes: training a neural network model based on actual data on a prediction target; and predicting an objective factor related to the prediction target based on explanatory factors related to the prediction target by the neural network model. The neural network model includes an input layer, an intermediate layer, and an output layer, and a coefficient of an activation function of the intermediate layer is larger than a coefficient of an activation function of the output layer.
1 . A method for prediction executed by an information processing device, the method comprising:
training a neural network model based on actual data on a prediction target; and
predicting an objective factor related to the prediction target based on explanatory factors related to the prediction target by the neural network model, wherein
the neural network model comprises an input layer, an intermediate layer, and an output layer, and
a coefficient of an activation function of the intermediate layer is larger than a coefficient of an activation function of the output layer.
2 . The method for prediction according to claim 1 , wherein
the activation function of the intermediate layer is a sigmoid function determined by below Mathematical Formula (1), and the activation function of the output layer is a sigmoid function determined by below Mathematical Formula (2), and
α1 and α2 satisfy α1>α2:
f
1
(
u
j
1
)
=
1
1
+
e
-
a
1
u
j
1
(
1
)
f
2
(
u
j
2
)
=
1
1
+
e
-
a
2
u
j
2
(
2
)
where
f
1
(
u
j
1
)
is the activation function of the intermediate layer;
α1 is the coefficient of the activation function of the intermediate layer;
u
j
1
is an input value to a j-th element of the intermediate layer;
f
2
(
u
j
2
)
is the activation function of the output layer,
α2 is the coefficient of the activation function of the output layer; and
u
j
2
is an input value to a j-th element of the output layer.
3 . The method for prediction according to claim 2 , wherein the prediction target includes a polycondensation reaction and an addition polymerization reaction.
4 . The method for prediction according to claim 2 , wherein a number of elements in the intermediate layer is 1.1 times or more and less than 6 times a number of the explanatory factors.
5 . The method for prediction according to claim 2 , wherein
a numerical value range of the explanatory factors input to the input layer is 0 or more and 1 or less, and
a numerical value range of the objective factor output from the output layer is 0.2 or more and 0.8 or less.
6 . The method for prediction according to claim 1 , wherein the prediction target includes a polycondensation reaction and an addition polymerization reaction.
7 . The method for prediction according to claim 1 , wherein a number of elements in the intermediate layer is 1.1 times or more and less than 6 times a number of the explanatory factors.
8 . The method for prediction according to claim 1 , wherein
a numerical value range of the explanatory factors input to the input layer is 0 or more and 1 or less, and
a numerical value range of the objective factor output from the output layer is 0.2 or more and 0.8 or less.
9 . An information processing device comprising:
a processor that
trains a neural network model based on actual data on a prediction target, and
predicts an objective factor related to the prediction target based on explanatory factors related to the prediction target by the neural network model, and
the neural network model comprises an input layer, an intermediate layer, and an output layer, and
a coefficient of an activation function of the intermediate layer is larger than a coefficient of an activation function of the output layer.
10 . A non-transitory computer-readable recording medium storing instructions executed by an information processing device that comprises a processor, the instructions causing the processor to execute:
training a neural network model based on actual data on a prediction target; and
predicting an objective factor related to the prediction target based on explanatory factors related to the prediction target by the neural network model, wherein
the neural network model comprises an input layer, an intermediate layer, and an output layer, and
a coefficient of an activation function of the intermediate layer is larger than a coefficient of an activation function of the output layer.