Search code examples
c#sentiment-analysismulticlass-classificationml.net

How to perform Binary Classification with many features using ML.Net

So I’ve been messing with ML.net and practicing using their examples and was wondering how I would train a model with the following input:

int, string, int, string, int, string, int, string, bool (Label)

and predict the binary label provided the other 8 features.

The first 3 strings are selections that could be converted to a lookup, and the fourth string is random text that needs more of a sentiment approach.

Any ideas on where to start / how to approach this? I’ve implemented using multiclass classification but I don’t think it makes sense since my output is binary.

Solution

  • Look at this sample from the ML.NET sample Github.

    You can use any of these Binary Classifiers for the lookups checkout the MapValueToKey Conversion, and for the additional fourth string you'll need to work with the Text Conversions, but there's nothing pretrained.

    I hope this helps!