Closed berisfu closed 6 years ago
whoa, interesting question.
you can certainly use auto_ml to predict multilabel classification problems, but unless your locations are a discrete number of buckets in the hundreds or less, that probably wouldn't work for you.
we don't support multilabel regression problems (and i don't actually know offhand of any library that does).
you might want to consider reframing your problem. you could do things like "distance_traveled" or "direction traveled" or "speed" or "arrived_on_time" or something else like that. but the problem as it's framed now probably won't work.
feel free to repoen if you want to chat more! i love talking about how to structure ml problems- that's become one of my favorite parts of the process now that auto_ml takes care of a lot of the more boring parts :)
Can it support multi label?for example,if Label is location(Longitude,Latitude),can I use auto_ml REGRESSOR to predict a location which is multi-label.