Hi I would like to propose a better implementation for 'test_indices':

We can remove the unneeded np.array casting:

Cleaner/New:
test_indices = list(set(range(len(texts))) - set(train_indices))

Old:
test_indices = np.array(list(set(range(len(texts))) - set(train_indices)))

What is the desired addition or change?

As discussed a bit in the PR. I suggested the idea of adding some data processing functionality to mlpack.

I compiled a list of features that can be implemented:

• head(): returns top 5 rows
• tail(): returns bottom 5 rows
• info(): Already implemented, I was thinking maybe

• What's wrong?

String representations of dataset objects are used for previewing their contents from the terminal. When converting a Dataset object to a string, we build a table using ascii characters. The current table has fixed width columns that do not take full advantage of the terminal real estate if the dataset only contains a few columns.

echo $dataset;

<img width="574" alt="Annotation

The PR JuliaData/CategoricalArrays.jl#310 means that an array with elements of type Symbol can no longer be wrapped as a CategoricalArray.

This means all MLJ documentation and test code that uses symbols in categorical data must be refactored to use strings instead.

These repos, at least, need checking/refactoring, in order of priority:

• MLJ

• [ ]

Hi @JavierAntoran @stratisMarkou,

First of all, thanks for making all of this code available - it's been great to look through!

Im currently spending some time trying to work through the Weight Uncertainty in Neural Networks in order to implement Bayes-by-Backprop. I was struggling to understand the difference between your implementation of `Bayes-by-Bac