search

yanyachen / rBayesianOptimization

Bayesian Optimization of Hyperparameters
84 stars 21 forks source link

how to specify bounds if the input is in a vector format #33

Open zjiang4 opened 6 years ago

zjiang4 commented 6 years ago

see the header....Many programmers do not provide the details about specifying the parameter format...that is really time consuming.

AnotherSamWilson commented 5 years ago

You would need to create multiple parameters that can be mapped from the scoring function to the vector. For instance, neuralnet has a parameter hidden which can be c(4,5) for a 2 hidden layer network. You would need to create 2 parameters firstLayer and secondLayer, create bounds for them and inside the scoring function assign them to the hidden parameter: hidden = c(firstLayer,secondLayer).