TY - GEN
T1 - A recommendation method based on contents and user feedback
AU - Kim, So Ryoung
AU - Choi, Sang Min
AU - Kim, Lae Hyun
AU - Han, Yo Sub
PY - 2012
Y1 - 2012
N2 - Nowadays, user is provided with many contents, which the previous search engines failed to find, thanks to various recommendation systems. These recommendation algorithms are usually carried out using collaborating filtering algorithm, which predicts user's preference, or contents based algorithm, which calculates on the basis of the similarity between contents. In addition to the above algorithms, many algorithms using user's context have been recently developed. Based on the previous researches, this paper proposes a new system to categorize contents information into various factors and learn user's selection. First, we divide information of items into four types and make user preference pattern using each information type. The information types can express more various user preferences and user preference pattern can calmly deal with user preference. Then, we calculate the score for recommendation using user preference pattern. That is, our system is constructed on these three modules: item analyzing module, user pattern analyzing module and recommendation score module. Lastly, we provide entire system flow to show how they work.
AB - Nowadays, user is provided with many contents, which the previous search engines failed to find, thanks to various recommendation systems. These recommendation algorithms are usually carried out using collaborating filtering algorithm, which predicts user's preference, or contents based algorithm, which calculates on the basis of the similarity between contents. In addition to the above algorithms, many algorithms using user's context have been recently developed. Based on the previous researches, this paper proposes a new system to categorize contents information into various factors and learn user's selection. First, we divide information of items into four types and make user preference pattern using each information type. The information types can express more various user preferences and user preference pattern can calmly deal with user preference. Then, we calculate the score for recommendation using user preference pattern. That is, our system is constructed on these three modules: item analyzing module, user pattern analyzing module and recommendation score module. Lastly, we provide entire system flow to show how they work.
UR - http://www.scopus.com/inward/record.url?scp=84883215644&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84883215644&partnerID=8YFLogxK
M3 - Conference contribution
AN - SCOPUS:84883215644
SN - 9781612081779
T3 - ACHI 2012 - 5th International Conference on Advances in Computer-Human Interactions
SP - 251
EP - 255
BT - ACHI 2012 - 5th International Conference on Advances in Computer-Human Interactions
T2 - 5th International Conference on Advances in Computer-Human Interactions, ACHI 2012
Y2 - 30 January 2012 through 4 February 2012
ER -