What is the article recommender about?
"You might be interested in ..." aka NUS Libraries article recommender (beta) is an AI-based system that aims to identify relevant and/or interesting library items to users based on usage data and content analysis of library resources.
The list of recommendations can help to broaden users’ initial search query and discover other resources by suggesting potentially relevant and/or interesting articles not picked up by the conventional search. Simply click on the links shown in the Article Recommender to access the specific resources (you may be prompted to login for subscribed resources).
How does the recommender work?
The recommender uses Machine Learning (ML) models and is trained based on usage data for library resources. It utilises algorithms such as collaborative filtering and content-based filtering to select and make the recommendations.
What kind of usage data does the recommender use?
The recommender uses transaction data from library physical collections and e-resources, for example, borrowing information and download logs of electronic articles. Besides usage data, it also uses library collections’ metadata.
Why do you store my transaction details?
The information is collected for security and business purposes to detect errant users, evaluate collection usage, improve services and provide user solutions such as the implementation of the article recommender system. These purposes are covered under the NUS Personal Data Notice for Staff/Students.
How is my privacy protected?
The raw data is stored in secured NUS facility and managed in accordance with the NUS Data Management Policy (DMP).
Will it replace the library search features and databases?
No, it is meant to complement the search features and databases. These existing search tools are good at showing you results based on keywords or when you know exactly what you are looking for. In contrast, the recommender makes recommendation by looking at similar items and/or usage by other users like you, balancing qualities such as relevance, appeal and serendipity.
For NUS Students and Staff, if you want to receive weekly email recommendations based on your profile in your NUS email, click on the Subscribe option and enter your NUSNET credentials when prompted.
You may unsubscribe anytime by clicking on the same button to uncheck the option.