MMR (Maximum Marginal Relevance) is an extractive summarization that was introduced by Jaime Carbonell and Jade Goldstein. http://repository.cmu.edu/cgi/viewcontent.cgi?article=1330&context=compsci
MMR aims to obtain the most relevance sentences by scoring whole sentences in the document. The MMR criterion strives to reduce redudancy while maintaining the content relevance in re-ranking retireved sentences
- This code is implemented in Python
- Since I built it for Bahasa Indonesia, I use Sastrawi libary for stemming
pip install Sastrawi - The list of stopwords is in Bahasa Indonesia (as provided in this folder)
- It also use "sklearn libarary". Please install it by
pip install sklearn - And this one
pip install termcolor
- In you terminal type it as
python mmr.py [Document.txt]
