http://delivery.acm.org/10.1145/1510000/1502702/p367-hu.pdf?key1=1502702&key2=9809528621&coll=ACM&dl=ACM&CFID=76752576&CFTOKEN=55465958
Authors:
Rong Hu - Swiss Federal Institute of Technology (EPFL)
Perl Pu - Swiss Federal Institute of Technology (EPFL)
This paper explained the differences between a ratings based approach to suggesting content and a personality quiz based approach to suggesting content to users. A ratings based approach will rely on the user rating content that they had already seen. For example if a user has seen movie X, they might rank movie X on a slider scale from 1 to 5. The system will then react to the rating and attempt to show more or less movies similar to movie X. In a personality based quiz method, the user will be asked many questions based on their personal likes or dislikes and suggestions will be provided based on these responses.
The authors chose two websites; MovieLens as the representative for the ratings based system, and WhatTorrent as the representative of the personality-based quiz approach. They took these two sites and provided a similar interface into both sites. Users were then put through the process on the sites and then took a survey on the users opinions about the suggestions given.
The researchers found that the personality-based quiz method generally yielded superior results. The personality-based approach saw great strides in how interested the user was in the results(see below). What was interesting to me however was that both system showed the same level of willingness to purchase. The personality based system did see a stronger desire to show to friends however so from a business stand point the personality based system should still be seen as being superior.
I really liked this idea. Now days there is so much content out there and I don't have time to sift through it all like I did back in High School so anything that helps me find content quickly is a positive for me. I would like to see more work being done on this and hopefully it will result in better ratings systems.
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment