RETRACTED: Implementation of Key Technologies for a Healthy Food Culture Recommendation System Using Internet of Things
Table 1
Advantages and disadvantages of recommendation algorithms.
Recommendation algorithm
Advantage
Shortcoming
Content-based recommendation algorithm
No need for users to access historical data, no storage in new objects and sparse problems, with mature classification learning support
Limited by feature extraction methods, there are new user questions and poor scalability, and training requires a large amount of data and high repetition rate of recommendation results
Collaborative filtering recommendation algorithm
Without user rating data, it is easy to find users’ new interest points, and the increase of the number of users can improve the recommendation effect
There are sparse and cold-start issues, and poor scalability, and the quality of recommendations is limited by historical data sets
Combined recommendation algorithm
The combination of the two methods effectively improves the recommendation effect
There are some challenges and difficulties in processing multi-source heterogeneous auxiliary information such as text and image