jnanliu / TopicModel

Implementation of a variety of topic models, include plsa, lda and so on

TopicModel

Simple implementation of some topic models, such as plsa, lda and so on

TODO

• plsa
• lda
• ...

Dependencies

• Python 3.6.2
• numpy 1.19.5
• jieba 0.42.1
• Scrapy 2.5.1 for scrawl data

All code only runs on Windows. I'm not sure if it can run on other platforms

Data

I crawled some articles from bilibili.

stopwords list is from stopwords

Train

You can run as follow

python main.py \
data_dir=data/bilibiliarticle \
number_of_topics=100 
max_iters=100

Note

• The code are not well tested, so it may contain bugs.

Reference

• http://blog.tomtung.com/2011/10/em-algorithm
• http://blog.tomtung.com/2011/10/plsa
• https://github.com/hitalex/PLSA/