Adding a folder "handeul_codes"

[handeulson]

In requirements.txt, I only added "scikit-learn". I created a simple machine learning code using "svm".

Test size was 0.2, and the accuracy was 47.22 %.

I also added numpy files.zip which I used for.

[ColinMoldenhauer]

Looks pretty good to me, I would suggest some minor changes:

• remove the data from your pull request

optional:

• use TreeClassifPreprocessedDataset instead of your loop solution

[ColinMoldenhauer]

• potentially add a seed so that the experiment is repeatable

[ColinMoldenhauer]

I personally would try to avoid adding data to git, because it will slow down git.

[ColinMoldenhauer]

As mentioned in Chris' pull request, please make sure the data is also not in your git history anymore by using git rebase (see this link)

[ColinMoldenhauer]

This probably works well, you could also use os.listdir(data_dir) to get all files directly, so that you don't need a while True: block.

Alternatively, you could make use of the new dataset class TreeClassifPreprocessedDataset, which already implements the logic of mapping files to class labels. So you could probably use something like

ds = TreeClassifPreprocessedDataset(data_dir)
for data_, label_ in ds:
    data.append(data_)
    labels.append(label_)