A movie recommendation engine using the Steamlit library for webgui. Recommendations are based off of a score calcuated as:
var1
= Users who liked (rated >= 4 stars) thegiven movie
(the movie that was searched for)var1_recs
= The other movies in our database that these users also likedperc1_recs
= Percent ofvar1
users that liked each of the recommendations, threshold of at least 10%var2
= Users who liked each of thevar1_recs
perc2_recs
= Percent ofvar2
users that liked each of the recommendations, no thresholdscore
=perc1_recs/perc2_recs
, or the share thatvar1
users make up of all the users that liked each movie
Searching for toy story
will give back Wreck-It Ralph (2012)
as the top recommendation, with an average user rating of 3.7 stars and a score of 5.07. The score says that 5x as many users liked Wreck-It Ralph
if they also liked Toy Story 3
.
A higher score should be more likely to reflect the unique likes of users who also liked the given movie