Q$\mathrm{.}1$: TFIDF and keyword$-$based methods a$.$ What is TFIDF? What is significance of IDF here? Explain role of IDF in classification with one real life example. Give an example where IDF may not make sense while classifying documents. b$.$ Explain how can you apply TFIDF to classify product reviews? $($Here each post has $20$ words$).$ Give Pseudo code for the same. c$.$ Discuss an algorithm to rank these reviews for different new customers based on their personalized parameters. $($Make necessary assumptions$)$ d$.$ Discuss an algorithm where you combine different reviews to provide a single personalized review of the product for a particular customer.

Q$\mathrm{.}2$: Topic modeling and PCA a$.$ What is topic modeling? Can textual artifact have multiple topics? Discuss two algorithms for topic modelling? b$.$ What are context pointers? Discuss their significance in topic modelling with one real life example. c$.$ What is PCA? Discuss PCA with one detailed example. d$.$ You got a set of news and now you want to cluster them and you use topic modelling as a first step. How will you do it$?$ e$.$ Discuss Implementation of topic modelling as hierarchical clustering. Explain different performance measures for your algorithm along with selection criteria.