PyAI 2.12

A generic python machine learning framework designed to be flexible and easy to use. It is built upon scikit-learn, numpy, scipy, and some custom written algorithms.

New in version 2.x:

Complete rewrite of main codebase in order to ease in adding new algorithms and much cleaner code

To test it, simply run:

import PyAI

PyAI.test()


The main object in the library is the Brain class (PyAI.Brain). With it you access all of the features in the framework.

brain = PyAI.Brain(x_data=data, y_labels=labels, y_data=reg_data)

| This brain object has 2 modes of operation: classification and regression.
| If you wish to perform classification (discrete) prediction, use the y_labels attribute
| If you wish to perform regression (continuous) prediction, use the y_data attribute
Or you can also provide both

| Then, you must initialize one of the algorithms available by performing:

brain.init_XXX()
# For example
brain.init_clustering(n_clusters=5)

Currently, the available algorithms are

- clustering
- neighbors
- svm
- gmm
- naive_bayes

Then you can apply any number of prediction methods in order to predict using the models

brain.predict_xxx_yyy
# For example
brain.predict_cluster_labels(test_data)
brain.predict_svm_data(test_data)

| The xxx must match on of the algorithms that you have initialized
| The yyy can either be 'labels' or 'data' for classification and regression respectively