What is the RBF kernel good for?

It lets an SVM draw curved boundaries instead of only straight lines, so it can separate two groups that no straight line could split apart.

How do I check an RBF SVM classifier?

Confirm there's one predicted label per row of new data. Beyond shape, measure accuracy on data you understand rather than trusting the array at a glance.

What libraries does this problem use?

Python, numpy, and scikit-learn's SVM with the RBF kernel. You train only on the training data and rely on scikit-learn rather than writing a kernel function yourself.

classify with a kernel SVM

Your task

An SVM finds the boundary that leaves the widest gap between two groups. With an RBF kernel it can draw curved boundaries, not just straight ones, so it can separate data that's tangled together. Complete svm_predict(X_train, y_train, X_test). Train scikit-learn's RBF-kernel SVM and return the predicted label for each row of X_test as an array. Example Two groups that no straight line can separate get split by a curved boundary.

Ground rules

•X_train and X_test are 2-D float arrays; y_train is a label array.
•use the RBF kernel.

Not required

•hyperparameter search over kernel type, C, or gamma.
•implementing a kernel function yourself.

▶About this challenge

This challenge asks you to complete a function that trains scikit-learn's SVM with the RBF kernel and returns a predicted label for each row of new data as an array. It takes training features and labels and the new data to predict, training only on the training data. It practices kernel classification — using a curved boundary to separate groups that no straight line can split.

An SVM finds the boundary that leaves the widest gap between two groups, and with an RBF kernel it can draw curved boundaries to separate data that's tangled together. This problem trains an RBF-kernel SVM with scikit-learn and verifies the predictions against real numbers rather than trusting a prediction array that merely looks the right shape.

What is the RBF kernel good for?: It lets an SVM draw curved boundaries instead of only straight lines, so it can separate two groups that no straight line could split apart.
How do I check an RBF SVM classifier?: Confirm there's one predicted label per row of new data. Beyond shape, measure accuracy on data you understand rather than trusting the array at a glance.
What libraries does this problem use?: Python, numpy, and scikit-learn's SVM with the RBF kernel. You train only on the training data and rely on scikit-learn rather than writing a kernel function yourself.