One-hot multinomial logistic regression

Quick Start

Installation

• To install ohmlr on your computer using pip, execute

  pip install ohmlr

• Test out ohmlr in Python:

  import ohmlr
  import numpy as np
  
  # create model and generate data
  n_features = 16
  n_x_classes = np.random.randint(2, 10, size=n_features)
  n_y_classes = 8
  model = ohmlr.ohmlr().random(n_features, n_x_classes, n_y_classes)
  x, y = model.generate_data(n_samples=1000)
  
  # fit and score model
  model.fit(x, y)
  print(model.score(x, y))

Links

Online documentation:

http://joepatmckenna.github.io/ohmlr

Source code repository:

https://github.com/joepatmckenna/ohmlr

Python package index:

https://pypi.python.org/pypi/ohmlr