ML Ecosystem in PHP

Learning examples

These examples will help you understand how you can use ML in PHP. They are not full-fledged applications, but they will help you understand the basics of working with ML in PHP.

A typical scenario: you have features from a database, you want to quickly train a model for classification or regression, save it, and use it in runtime without external services.
Let's consider a simple and illustrative example. In it, we train a k-nearest neighbors (k-NN) classifier on a small set of points, each of which belongs to one of two classes — $a$ or $b$. After training, the model must determine which class a new point belongs to. We specify the training set as coordinates on a plane and the corresponding class labels. For the point $[3, 2]$, the algorithm returns class b because its nearest neighbors in the training set belong to this class.

Rubix supports classification, regression, clustering, and working with datasets as first-class objects. Let's see how this looks in practice. Let's say we have data for binary classification.
The code below also trains a k-nearest neighbors (k-NN) classifier, but this time on height and weight data with gender labels, and then predicts the label $M$ for a new person. The model returns $[172, 68]$ for the parameters, since most of the 3 nearest neighbors have this label.

TransformersPHP gives you a simple pipeline API for tasks like sentiment analysis, text classification or semantic comparison. The snippet below runs a sentiment model on two short phrases and shows the resulting label and score.