A movie streaming platform uses an item$-$based collaborative filtering approach for their recommendation system. They notice that the system is recommending mostly popular movies and not capturing individual user preferences well. What could be a possible reason for this issue?

Group of answer choices

The platform has a diverse movie catalog, making it difficult to find similar items

The similarity metric used to calculate item similarities is not appropriate

The platform has a limited number of users, leading to a sparse user$-$item matrix

The system is not updating the item similarities frequently enough