Abstractive and Extractive Methods For Automatic Summerization
This report contains:
- The first one is a really simple model where you just give scores to words and sentences based on how frequent they are without having meaning behind words by not using any language models.
- The second one using Text Rank Algorithm and Word embeddings, TextRank is an extractive and unsupervised text summarization technique.
- Using sequence to sequence Model and attention model to get an understand of articles and focusing on the important aspects of it to abstract and useful summerization
ps: I could not upload the dataset and language models used due to their large size, however they can be found here: https://drive.google.com/open?id=17um71iYzdUMTDnlalpMZvzgcpfZkgw5U