Tensor LDA

tensor-lda is a LDA (Latent Dirichlet Allocation) implementation that use tensor decomposition method to estimate LDA parameters. It follows scikit-learn's API and can be used as scikit-learn's extension.

HTML Documentation - https://chyikwei.github.io/tensor-lda/

Install:

# clone repoisitory
git clone [email protected]:chyikwei/tensor-lda.git
cd tensor-lda

# install numpy & scipy
pip install -r requirements.txt
pip install .

Example:

The usage is the same as scikit-learn's LDA model.

Here is an example that extract topics from 20 news grooup dataset.

Running Test:

python setup.py test