d$)$ Using TF$-$IDF, a movie can be represented by a $5-$dimension vector of TF$-$IDF values

of $5$ tags in the dictionary. Compute the vector for each movie.

e$)$ Similarly, any given query can be represented by a $5-$dimension vector. Compute the

vector for the query $$Action$.$

f$)$ Using document vectors, we can compute the relevance score for each document to a

given query using cosine similarity between the document vector and the query vector. A

retrieval result of a query is the ranking of the movies in the decreasing order of

relevance scores. What is the result of the query $$Action$?$

g$)$ If each movie has an additional tag $$Hollywood$,$ what is the result of the query

$$Action$?$ What is the result of the query $$Hollywood$?$ And what is the result of the query

$$Hollywood Action$?$

h$)$ If the appearance order of tags in each movie matters, how do you propose to solve

the retrieval problem? What is the result of the query $$Action$$ now?