@sylvaticus Its me again!.
I have installed the under-development (in fact, the “stable package”) package BetaML. The example runs smoothly!. Now, I have been trying to use the old version and I was struggling to get a good prediction over a simple training data set of my own invention. I generated.
The training data set was this:
xtrain = pi*rand(10000,5);
ytrain = sin.(xtrain) + 0.5 * cos.(xtrain);
So, actually, it represents a combination of basic trigonometric functions. I was then trying to test it with a subset like the one below…
xtest = pi*rand(1,5);
ytest = sin.(xtest) + 0.5 * cos.(xtest);
and I get
in green, the training data.
in red, the testing data,
in blue: the prediction. (obtained with Predict(myfnn,xtest))
As you can see, its not good!. I am probably doing something wrong…
My question is, can yo run this example using the new version, so I can see how it is used?.
Thanks in advance!
Cheers and I hope this message finds you well