Table 8 Hyperparameters of the baseline models.
From: Time series transformer for tourism demand forecasting
Model | Dataset | |
|---|---|---|
Jiuzhaigou Valley | Siguniang Mountain | |
(seasonal) Naive | S = 1 for single step, S = 7 for multistep | |
ARIMA | p = 4, d = 1, q = 4 | p = 4, d = 1, q = 3 |
ARIMAX | p = 4, d = 1, q = 4 | p = 4, d = 1, q = 3 |
SARIMA | p = 4, d = 1, q = 4, P = 1, D = 0, Q = 1, S = 7 | p = 4, d = 1, q = 3, P = 0, D = 1, Q = 1, S = 7 |
SARIMAX | p = 4, d = 1, q = 4; P = 1, D = 0, Q = 1, S = 7 | p = 4, d = 1, q = 3, P = 0, D = 1, Q = 1, S = 7 |
SVR | kernel = rbf, gamma = 0.01, C = 1, epsilon = 1e-5 | kernel = rbf, gamma = 0.001, C = 10, epsilon = 1e-5 |
k-NN | neighbors = 5, leaf size = 5, p = 2 | neighbors = 10, leaf size = 5, p = 1 |
ANN | hidden layers = 2, learning rare = 1e-2, hidden size = 128 | |
LSTM | hidden size = 128, learning rare = 1e-2 | |
