[Retracted] Research on Data Mining Algorithm of Associated User Network Based on Multi-Information Fusion
Table 3
Comparison of UserCF and ItemCF.
UserCF
ltemCF
Performance
Applicable to scenarios with a small number of users. It is very expensive to calculate the user similarity matrix when there are many users.
Applicable to the scenario where the number of items is significantly smaller than the number of users, and it costs a lot to calculate the item similarity matrix when there are many items.
Field
It applies to areas with strong timeliness and less obvious user interest.
It is suitable for the fields with rich long tail items and strong personalized needs of users.
Real-time performance
New user behavior does not necessarily result in immediate changes in recommendation results.
New user behavior is bound to result in real-time changes in recommendation results.
Cold start
New users cannot make personalized recommendations immediately after they have behaviors for a few items.
By acting on an item, a new user can recommend other items related to that item.