Customized ranking by user preference using LRR model

Latent Aspect Rating Analysis (LARA);Recommendation system

Chiang, Bo-Yu

关键词: Latent Aspect Rating Analysis (LARA);    Recommendation system;   

Others  :  https://www.ideals.illinois.edu/bitstream/handle/2142/78480/CHIANG-THESIS-2015.pdf?sequence=1&isAllowed=y

美国|英语

来源: The Illinois Digital Environment for Access to Learning and Scholarship

【 摘 要 】

In this thesis, we proposed a customized ranking system that can rank all the entities given a specific user preference. Rank entities by user’s preference is an inevitable strategy of saving user’s time browsing and extracting usefulinformation from Internet. Modern websites always rank these entities by a single numeric value computed by averaging overall rating, but this ranking scheme is of limited use to users.With di↵erent aspect preference, it is obvious that the restaurants rankingshould be di↵erent based on their famous features, e.g., service, environment, price. We used the LRR (Latent Rating Regression) model to aggregate restaurants aspect score and proposed two ranking approaches. The experimentresults show that the two ranking approaches are both better than thebaseline ranking approach.