qWiki search helps user to gain knowledge about a topic in a jiffy. User enters topic and number of sentences he wants, as a result the user receives the relevant and latest sentences about the topic.
We find relevant wikipedia pages about the topic entered by the user. Next, sentences containing dates are extracted from the pages. tf-idf vectorization is computed for these sentences. This is then given as input for non-negative matrix factorization.
Non - Negative Matrix Factorization (NMF) splits the document-word matrix (tfidf) into document-topic matrix (W) and topic-word matrix(H). Sentences giving highest score are then extracted from W and latest sentences are shown in chronological order.
re 2.2.1
nltk 3.2.5
np 1.19.5
sklearn 0.22.2.post1
wikipedia
punkt (nltk.download('punkt')
Run the Google colaboratory notebook which explains the individual steps.