laserwave / lda_variational_em

a python implementation of latent dirichlet allocation(lda) using variational EM algorithm

LDA (Latent Dirichlet Allocation)

This is a python implementation of LDA using variational EM algorithm.

The following picture shows the top 10 words in the ten topics generated by this algorithm over 16 sentences about one piece on wikipedia.

The code contains both the training of the model and predicting topic of new documents.

The following picture shows the top 10 words in the 30 topics (set K = 30) generated by this algorithm over 5000 chinese sina social news.

Author

• ZhikaiZhang
• Email [e-mail]
• Blog http://zhikaizhang.cn
• 自然语言处理之LDA II-Variational EM实现