dc.contributor.advisor | Razzaghi, Talayeh | |
dc.contributor.author | Welborn, Thomas | |
dc.date.accessioned | 2021-12-13T16:04:30Z | |
dc.date.available | 2021-12-13T16:04:30Z | |
dc.date.issued | 2021-12 | |
dc.identifier.uri | https://hdl.handle.net/11244/332304 | |
dc.description.abstract | As the amount of information users interact with every day continue to grow, filtering it for useful information is increasingly important. One of the most useful tools for this task are recommender systems (RS). These look at past products the user has interacted with and recommends similar products. However, these suffer from a major issue, cold-start, in which there is difficulty in producing recommendations for new users. One of the suggested techniques for mitigating the cold-start issue is the use of trust data. By using the relationships between users such as friendships on social media or following reviewers of movies the recommender system can recommend products that the user’s friend would rate highly as well.
We extend previous trust models by applying a One-Class Support Vector Machine model to the known trust relations and predicting distrust relations among users. This is shown to improve the predictions movie ratings in some circumstances. | en_US |
dc.language | en_US | en_US |
dc.subject | Statistics. | en_US |
dc.subject | Data Science. | en_US |
dc.subject | Recommender Systems. | en_US |
dc.subject | Trust-Aware. | en_US |
dc.title | One-class Support Vector Machines Approach for Trust-Aware Recommendation Systems | en_US |
dc.contributor.committeeMember | Hougen, Dean | |
dc.contributor.committeeMember | Nicholson, Charles | |
dc.date.manuscript | 2021-12-02 | |
dc.thesis.degree | Master of Science | en_US |
ou.group | Gallogly College of Engineering | en_US |