Using River for online learning

In this section, you will discover the River library for online learning. River is a Python library that is made specifically for online machine learning. Its code base is a result of the combination of the creme and the scikit-multiflow libraries. The goal of River is to become the go-to library for machine learning on streaming data.

In this example, you'll see how to train an online model on a well-known dataset. The steps that you will take throughout this example are the following:

1. Import the data.
2. Reclassify the data to obtain a binary classification problem.
3. Fit an online model for binary classification.
4. Improve the model evaluation using a train-test split.
5. Fit an online multiclass classification model using one-vs-rest.

Training an online model with River

For this example, you will use the well-known iris dataset. You can download it from the UCI machine learning repository, but you can also use the...