irmlearn is a algorithms for Infinite Relational Model.
This contains these techniques.
- IRM
- Poisson IRM
The required dependencies to use irmlearn are,
- scikit-learn
- numpy
- scipy
You also need matplotlib, seaborn to run the demo and pytest to run the tests.
pip install irmlearn
We have posted a usage example in the demo folder.
from irmlearn import IRM
alpha = 1.5
a = 0.1
b = 0.1
max_iter = 300
model = IRM(alpha, a, b, max_iter, verbose=True, use_best_iter=True)
model.fit(X)
from irmlearn import PoissonIRM
alpha = 0.5
a = 5
b = 5
max_iter = 300
model = PoissonIRM(alpha, a, b, max_iter, verbose=True, use_best_iter=True)
model.fit(X)
This code is licensed under MIT License.
python setup.py test