Location:Home > Engineering science > Computer Science > Applied Computer Technology > Research on Personalized Recommendation Based on Probabilistic Relational Models

Research on Personalized Recommendation Based on Probabilistic Relational Models

Downloads: []
Tutor: LiuDaYou
School: Jilin University
Course: Applied Computer Technology
Keywords: Personalized Recommendation,Likelihood relational model,Statistical Relational L
CLC: TP393.092
Type: Master's thesis
Year:  2008
Facebook Google+ Email Gmail Evernote LinkedIn Twitter Addthis

not access Image Error Other errors

In recent years , personalized recommendation system for universal application in the e-commerce site . However, the rapid growth of the number of effective online information and the types of goods on the recommended system requirements , personalized recommendation algorithm cold start the sparse user evaluation data and recommended real-time problems to be solved . Likelihood statistical relational learning Relational Model (PRM) personalized recommendation based on project characteristics , user characteristics and collaborative filtering technology , put forward three for personalized recommendation likelihood relational model , and these three Recommended model weighted switching is proposed based on the combination of the PRM recommended model . Above model makes full use of project information, user information and user ratings data on the project model , with a first-order features , the recommended process does not depend on the specific user and project to solve the problem of the the recommendation algorithm cold start ; Once the model structure is completed , recommend higher efficiency , and can ensure the real-time requirements of the algorithm . MovieLens datasets verify the recommended performance of the algorithm , the average absolute deviation (MAE) is relatively low . Based PRM recommended model is applied to the network of educational resources management system ( NERMS ) , designed and personalized recommendation subsystem to provide users with the resources initiative recommendation service , to achieve the expected results .
Dissertation URL:
Related Dissertations
Last updated
Sponsored Links
Home |About Us| Contact Us| Feedback| Privacy | copyright | Back to top