Author: Siddharth Patel
This repo is part of my work in NLP to find important phrases from given statement.
This repo has complete code but it doesn't give proper result as I am not having enough data to create vocab. As of now, this work uses en_core_web_lg
from Spacy
to find similarity in text.
This can be better trained with Word2Vec
with gensim
to get far better results for this task given that we have enough data to train our system.
In case you find good dataset for task, do let me know on Twitter @Siddharth1India and I can add you as contributor.
Given good data, this project can be extended to many important usecases in real-world like extracting information from various data sources which actually matters rather than spending time reading useless text data.
For any further queries, please open issue, I will respond within hours.