Made SVM multiclass

[DanielLacina]

I split the SVM classification into multiple binary classifications using the one for all strategy.

[DanielLacina]

I probably messed a ton of things up. Im new to open source.

[DanielLacina]

Out of 50-100 iterations of random sampling, the lowest accuracy score on the test set was 0.8666667

[DanielLacina]

It's extremely slow relative to scikit-learn though

[Mec-iS]

Thank you! I will have time to take a look at this later this week or next week 🌈

[DanielLacina]

Alright. Appreciate you notifying me.

[Mec-iS]

np. I have enabled the tests runners in the meanwhile.

[DanielLacina]

Ima fix the test issues with this PR and then work on other issues.

[Mec-iS]

Something to start from:

Please measure performance before and after the change. See the workbenches https://github.com/smartcorelib/smartcore-benches
you can add a bench with the thing you are working on.

to use an array in smartcore you need to use the provided zero-copy structures like Array1 and Array2, not Vec. For example at line 152: Array1<TY> should become Array2<TY>