In my prpject , I use unsupervise model train in data on the site : soha.com, dantri.com, kenh14.com, ...
I using Doc2vec, Word2vec, LDA, Autoencoder are the unsuperivse model for representation sentence to vectorizer. After, I use Pagerank and K-mean for select sentence from graph or cluster. All sentence selected will be throught MMR or plMMR for select sentence for create summary of document begin input.
For requestment :
- python >= 3.5
- module using in file requestment.txt
Run :