I am struggling to understand the point of supervised ML models in real world scenarios

Question

Sorry for maybe a stupid question, but I can't seem to find any explanation of it online. If supervised machine learning only works on labeled datasets - you can't use it to predict a value of unlabelled data after the model is already trained?

And if that is true, how could you possibly use those models in real life scenarios? For example you write and train a classifier to predict the age group of the user, can you in any way use that created model for actual prediction of an unlabelled entry? And if not, what is the point of building this kind of model?

Thank you!

Answer 1

Supervised means that the training stage is supervised and requires labels. It does not mean that you need labels during inference.

Here is small example using a Random Forest classifier with scikit learn (source):

>>> from sklearn.ensemble import RandomForestClassifier
>>> from sklearn.datasets import make_classification
>>> X, y = make_classification(n_samples=1000, n_features=4,
...                            n_informative=2, n_redundant=0,
...                            random_state=0, shuffle=False)
>>> clf = RandomForestClassifier(max_depth=2, random_state=0)
>>> clf.fit(X, y)
RandomForestClassifier(...)
>>> print(clf.predict([[0, 0, 0, 0]]))
[1]

As you can see, training the model takes in labels (clf.fit(X, y) where y are the labels) but during inference the model runs the prediction for an unseen datapoint with no label (print(clf.predict([[0, 0, 0, 0]])) where [[0, 0, 0, 0]] is a a new datapoint which is classified as belonging to class 1).

In contrast, unsupervised ML does not require labels during training. This blog post provides some further explanations and examples.