A Self-Training algorithm based on Fuzzy C-Means clustering

Usage

1. Create a dataloader for your dataset (inherit from AbstractDataLoader)
2. Pick a classifier model, optionally a kernel approximation algorithm
3. Initialize the SemiSupervisedModel class instance
4. Increase the labeled part of the dataset by calling the increase_labeled method
5. Fit (fit) and predict (predict)