Implementations of a couple of semi-supervised machine learning algorithms:
- A Positive-Unlabelled adaptor: works in the setting where only positive labels are observed. Converts a supervised classifier to a semi-supervised one.
- Semi-supervised Gaussian Mixture Model: semi-supervised adaptation of the (unsupervised) GMM method. Works for an arbitrary number of incomplete label classes.
More information about using the code can be found in the Notebook, and the theory underpinning these algorithms can be found here