I'm using Gpy to train a gaussian process regression model. The dimension for the input data is 4, the corresponding outputs' dimension shall be 3. I tried to just build one model for the outputs, and the model doesn't work at all.
I tried to build separate GPR model for the output which has a dimension of 1728 by 3. And it doesn't work.
import GPy
### x_train.shape: 1728*4
### y_train.shape: 1728*3
n_dim=x_train.shape[1]
kernel=GPy.kern.RBF(input_dim=n_dim,variance = 0.5)
model=GPy.models.GPRegression(x_train,y_train,kernel)
model.optimize()
y_test=model.predict(x_test)[0]
