Hi, thank you for your very sincere help, and I spent several days trying to understand your suggestion in my case for I’m a novice.
##Tile and reshape your input matrix, (88,2) → (35200,2)
X_tiled = np.tile(X,[1,1,2]).reshape([-1,2])
I think it should be:
X_tiled = np.tile(X,[1,1,400]).reshape([-1,2])
Hope I get what you mean. And when I use the predict function, enter two parameters, I will get one output value.
What I want to say is, my previous description of the observations may be a bit vague, I originally thought that entering two values can get 400 values, like the Gaussian process model in sklearn object, because these 400 values are in order, they are not randomly sample from a certain distribution, each has different sequence and position. (I don’t know if I have expressed it clearly). So I hope to achieve the effect of establishing a gp model for each measurement point.
All in all, thank you very, very much, I think I might find a way to process my training data, and use the method you suggested to achieve my goal.