English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

A typical benchmark dataset for recommender system (RecSys) evaluationconsists of user-item interactions generated on a platform within a timeperiod. The interaction generation mechanism partially explains why a userinteracts with (e.g.,like, purchase, rate) an item, and the context of when aparticular interaction happened. In this study, we conduct a meticulousanalysis on the MovieLens dataset and explain the potential impact on using thedataset for evaluating recommendation algorithms. We make a few main findingsfrom our analysis. First, there are significant differences in userinteractions at the different stages when a user interacts with the MovieLensplatform. The early interactions largely define the user portrait which affectthe subsequent interactions. Second, user interactions are highly affected bythe candidate movies that are recommended by the platform's internalrecommendation algorithm(s). Removal of interactions that happen nearer to thelast few interactions of a user leads to increasing difficulty in learning userpreference, thus deteriorating recommendation accuracy. Third, changing theorder of user interactions makes it more difficult for sequential algorithms tocapture the progressive interaction process. Based on these findings, wefurther discuss the discrepancy between the interaction generation mechanismthat is employed by the MovieLens system and that of typical real worldrecommendation scenarios. In summary, models that achieve excellentrecommendation accuracy on the MovieLens dataset may not demonstrate superiorperformance in practice for at least two kinds of differences: (i) thedifferences in the contexts of user-item interaction generation, and (ii) thedifferences in user knowledge about the item collections.

Our Model Achieves Excellent Performance on MovieLens: What Does it Mean?