Add spam dataset from Elements of Statistical Learning

[plato-12]

Add spam_dataset from Elements of Statistical Learning (GSoC-25 tasks)

Description

This PR adds a new dataset implementation for the spam email classification dataset from the "Elements of Statistical Learning" book. This dataset contains 4601 emails with 57 features and a binary classification indicating whether the email is spam or not.

Features

• Implementation of spam_dataset() function with documentation
• Test suite to verify dataset functionality
• Example vignette showing how to use the dataset for a classification task

Use Case

This dataset can be useful for binary classification exercises and demos, as it's a well-known dataset in the statistical learning community. It's relatively small (compared to image datasets) but provides a realistic classification problem.

Testing

• All tests for the new dataset pass
• Manual verification of the dataset loader and example usage has been performed

[dfalbel]

Hi @plato-12,

Can you direct your Pr to trochdatasets? https://github.com/mlverse/torchdatasets