RETRACTED: Personalized Recommendation System Design for Library Resources through Deep Belief Networks

Table 1

Comparison among different recommendation algorithms.

Recommendation algorithm	Advantage	Disadvantage
Content-based	The recommended results are intuitive and easy to interpret	Complex attributes of the user are difficult to handle
User statistics-based	No new user issues	Demographic information on users is difficult to obtain
Utility-based	More sensitive to changes in user preferences	Poor flexibility
Association rule-based	Easy to find new points of interest for users	Difficult and time-consuming to extract association rules