1 / 5
Showing trending or top items to a brand-new user with no history is a ___ fallback.
-
-
-
-
A popularity fallback gives reasonable recommendations to cold users until enough personal signal accumulates.
2 / 5
Recommending items by matching their attributes to a new item's attributes is ___ bootstrapping.
-
-
-
-
Content-based methods use item features (genre, tags) so a new item can be recommended before it has any interactions.
3 / 5
Asking new users to pick a few interests during signup gathers onboarding ___.
-
-
-
-
Explicit onboarding signals (preferred topics, sample ratings) jump-start personalisation despite no behavioural history.
4 / 5
Occasionally showing uncertain items to learn a new user's taste is called ___.
-
-
-
-
Exploration trades some short-term relevance to gather signal, balancing the explore/exploit trade-off for cold users.
5 / 5
The cold-start problem is hardest because the system lacks ___ for the new user or item.
-
-
-
-
With no interaction history, collaborative filtering has nothing to learn from, which is exactly what these strategies work around.