Tagaware Recommender Systems by Fusion of Collaborative Filtering

Tag-aware Recommender Systems by Fusion of Collaborative Filtering Algorithms (Cited by 5) Karen H. L. Tso-Sutter, Leandro Balby Marinho* and Lars Schmidt-Thieme Information Systems and Machine Learning Lab (ISMLL) University of Hildesheim Samelsonplatz 1, 31141 Hildesheim, Germany in Proceedings of 23 rd Annual ACM Symposium on Applied Computing (SAC'08), Fortaleza, Brazil. Presented by Jun-Ming Chen 2/27/2009

Outline n n Introduction Collaborative Filtering q q n n Extension with Tags Fusing User-based and Item-based Experiments Conclusions 2

Introduction n Tag (Action : tagging) q q n enabling people to easily add metadata to content Improve search mechanisms Collaborative tagging system folksonomies q Web services n q flickr, del. icio. us Advantage n n better structure the data for browsing provide personalized recommendations fitting the users’ interests 3

Introduction n Recommender Systems (RS) aim at predicting items or ratings (preference score, 1 -5) of items that the user are interested in. n To improve recommendation quality, metadata such as content information of items has typically been used as additional knowledge. n (attribute aware) RS q algorithms is typically attached to the items and is usually provided by domain experts n an item always has the same attributes among all users 4

Introduction n tags are provided by various users q q n not only associated to the items but also to the users act as additional background knowledge to improve RS algorithms With the increasing popularity of the collaborative tagging systems, tags could be interesting and useful information to enhance RS algorithms. 5

Introduction n In this paper, we propose q To integrate tags in recommender systems n n n by first extending the user-item matrix and An adapted fusion mechanism to capture the 3 -dimensional correlations between users, items and tags Empirical evaluations on three CF algorithms with real-life data set demonstrate that incorporating tags to our proposed approach provides promising and significant results 6

Collaborative Filtering n Recommender systems (RS) predict ratings of items or suggest a list of items that is unknown to the user q n Most recommender systems derive recommendations to a user q q n Two di�erent recommendation tasks n predicting the ratings, i. e. how much a given user will like a particular item n predicting the items, i. e. which N items a user will rate, buy or visit next (top. N) by using opinions from people who have alike tastes, called neighborhood, while concealing the real identity of the users neighborhood. The prevalent method in practice is Collaborative Filtering (CF) 7

Collaborative Filtering n Its idea is basically the nearest neighbor method n Given some user profiles, q n it predicts whether a user might be interested in a certain item, based on a section of other users or items in the database. There are in general two types of collaborative filtering: q q user-based item-based 8

l User-Based l Item-Based l l Cosine Probability 9

User-Based CF - Existing issues n Low Performance when with millions of users and items. (userbased filtering--KNN) n Low Prediction accuracy when with sparsity of data 2008 -9 -25 10

User-Based CF - Solution n Explore item-based collaborative filtering n n Method: Build user-item matrix identify relationships between different items use relationships to make recommendations n In contrast, item-based collaborative filtering q q analyzes historical transaction information to identify relations between items that are co-purchased, uses these relations to compute the similarities between items, and make recommendations based on item similarity. 2008 -9 -25 11

User-Based CF - Solution n It constructs a primitive model with significant amount of time, but makes recommendations in a very short time. n It then alleviates the problems of scalability and real- time performance that user-based CF faces. 2008 -9 -25 12

Collaborative Filtering n Notations 13

Collaborative Filtering n In user-based CF [19], recommendations are generated by q q considering solely the ratings of users on items, by computing the pairwise similarities between users, n e. g. , by means of vector similarity: w(u, v) Useru Userv [19] B. M. Sarwar, G. Karypis, J. A. Konstan, and J. Riedl. Analysis of recommendation algorithms for e-commerce. In Proceedings of the Second ACM Conference on Electronic Commerce (EC’ 00), pages 285– 295, 2000. (Cited by 634) 14

Collaborative Filtering - recommendations n In user-based CF, to derive the recommendations for a target user u, q n usually only similarities of the k most-similar users are selected (neighborhood – Nu). When predicting a rating of a given user u for an item i, the weighted sum of the other users are computed by: 15

Collaborative Filtering n A dualistic form of user-based CF is item-based CF [9], where similarities are computed between each pair of items. w(i, j) [9] M. Deshpande and G. Karypis. Item-based top-n recommendation algorithms. ACM Transactions on Information Systems. Springer-Verlag, 22/1, 2004. (Cited by 168) 16

Collaborative Filtering - recommendations n In the case of item-based CF, q n the prediction would be the average of the ratings of k most-similar items Ni rated by the given user u. In similar notation as user-based CF, q the prediction for a rating of a given user u for an item i u 17

Extension with Tags n < item, attribute > < user, item, tag > n We cope with this three dimensionalities by projecting it as three two-dimensional problem, q q q < user, tag > < item, tag > < user, item > 18

Extension with Tags n User tags, are tags that user u, uses to tag items and are viewed as items in the user-item matrix. n Item tags, are tags that describe an item, i, by users and play the role of users in the user-item matrix 19

Extension with Tags n Let: 20

Extension with Tags n To apply user- and item-based CF after the extension with tags, both CF algorithms have to be recomputed with the newly extended useritem matrix. n For user-based CF, the new user-item matrix , q Ruextend : = R + RTu, is represented in a U × Iextend matrix Set : Users x n (Items + user Tags) For item-based CF, the new user-item matrix, q Riextend : = R + RTi , is represented in a Uextend × I matrix Set : (Users + Item tags) x Items Ruextend : = R + RTu U × Iextend matrix 21

Fusing User-based and Itembased n Tags of an item, are descriptions of the item by one or more than one users. Thus, tags are not only attached to the item itself but also are depended on the user’s preference. n This suggests that a RS algorithm that is able to q n capture both user’s and item’s aspect of tags would eventually be a suitable choice. We have selected an existing algorithm developed by Wang et al. [21], which fuses the predictions of user- and item-based CF. [21] J. Wang, A. P. de Vries, and M. J. T. Reinders. Unifying user-based and item-based collaborative filtering approaches by similarity fusion. In SIGIR ’ 06: Proceedings of the 29 th annual international ACM SIGIR conference on Research and development in information retrieval, pages 501– 508, New York, NY, USA, 2006. ACM Press. (Cited by 49) 22

Fusing User-based and Itembased n The fusion of the user- and item-based predictions was done by q q computing the sum of the two conditional probabilities that are based on user- and item-based similarities, which are computed using standard user- and item-based CF. § A parameter, λ, is introduced to adjust the significance of the two predictions. j λ = 0. 4 23

Fusion for Predicting Item Problem n Most systems that use collaborative tagging do not contain rating information, q n Yet, the fusion method by Wang et al. [21] q q n i. e. only the occurrence, Ou, i ∈ {0, 1} does not consider tags. not the predicting item problem Thus, we propose a fusion algorithm that tackles the predicting item problem and also takes tags into account. [21] J. Wang, A. P. de Vries, and M. J. T. Reinders. Unifying user-based and item-based collaborative filtering approaches by similarity fusion. In SIGIR ’ 06: Proceedings of the 29 th annual international ACM SIGIR conference on Research and development in information retrieval, pages 501– 508, New York, NY, USA, 2006. ACM Press. 24

Fusion for Predicting Item Problem n For the predicting item problem in user-based CF, recommendations are a list of items that is ranked by decreasing frequency of occurrence in the ratings of his/her neighbors. n For item-based CF, the top. N recommendation suggested by [9] is to compute a list of items that is ranked by decreasing sum of the similarities of neighboring items, Ni, which have been rated by user u. n user-based being the frequency of items, and item-based the similarity of items [9] M. Deshpande and G. Karypis. Item-based top-n recommendation algorithms. ACM Transactions on Information Systems. Springer-Verlag, 22/1, 2004. (Cited by 168) 25

Fusion for Predicting Item Problem n To address the di�erent meanings of the values, we normalized the prediction lists to unity n The top. N combined prediction list is thus: 26

Fusion for Predicting Item Problem n We deem that this tag-fusion algorithm is a suitable tag-aware RS algorithm because q q n user tags and item tags provide extra indications of the user’s and item’s preferences, our adapted fusion approach then brings about both user and item aspects of the tags concurrently. In fact, our empirical analysis has shown that our tag-fusion RS approach provides promising results. 27

Experiments n July 2006 by crawling the Last. fm site q q q n 1853 items (artists), 2917 users and 2045 tags users preference information n songs the users listen to tags information n user-end tagging of artists, albums, and tracks of the music Appear at least 10 times in the tag assignments, q i. e. (user, resource, tag) triples Last. fm is a popular internet radio and music community website which allows user to tag the music. 28

Experiments n 10 -fold cross validation q n focuses on the item prediction problem, which is q n split the training data to validation data to optimize n the parameters λ(0. 4) and n k (20), the neighborhood size to predict a fixed number of top. N recommendations (N = 10) Suitable evaluation metrics are q Precision, Recall and F 1 Last. fm is a popular internet radio and music community website which allows user to tag the music. 29

Experiments n It can be seen that the fusion method, both with and without tags, significantly outperform the standard CF models. 30

Experiments n It is interesting to see that incorporating tags to the baseline models does not improve the recommendation quality at all, in contrast to the promising results of including tags in the fusion method. 31

Experiments n Applying user/item tags alone does not exploit the characteristic of tags correctly. q q n Hence, attaching tags to standard CF algorithms, does not improve the performance at all, the tags are then only seen as noise. By simply extending the standard CF algorithms with tags, it fails to denote the 3 -dimensional correlations between user, item and tag, q whereas the proposed fusion method has shown to be able to capture this relationship. 32

Conclusions n We have presented a generic method to include tags to standard CF algorithms such as user- and item-based CF. n We have found an approach that deals with the 3 -dimensional correlation between the users, items and tags by q q first applying our tag extension mechanism and then an fusion method which we have adapted from a predicting rating problem to predicting item problem. n Our empirical analysis has shown that the proposed adapted fusion method outperforms standard baseline models, especially with the incorporation of tags. n Moreover, our findings have suggested that our adapted fusion method has successfully captured the relationships between users, items and tags. 33

IWi. LL - 增進學生閱讀學習的 Keywords 活動 passion darkness obsession Catherine orphan lovely cover art Presented by Jun-Ming Chen (Jacky)

ox 小說 分享 Keyword 收穫 討論 Keyword Knowledge Keyword

When George was about six years old, he was made the wealthy master of a hatchet of which, like most little boys, he was extremely fond. He went about chopping everything that came his way. One day, as he wandered about the garden amusing himself by hacking his mother's pea sticks, he found a beautiful, young English cherry tree, of which his father was most proud. He tried the edge of his hatchet on the trunk of the tree and barked it so that it died. Some time after this, his father discovered what had happened to his favorite tree. He came into the house in great anger, and demanded to know who the mischievous person was who had cut away the bark. Nobody could tell him anything about it. Just then George, with his little hatchet, came into the room. "George, '' said his father, "do you know who has killed my beautiful little cherry tree yonder in the garden? I would not have taken five guineas for it!'' This was a hard question to answer, and for a moment George was staggered by it, but quickly recovering himself he cried: "I cannot tell a lie, father, you know I cannot tell a lie! I did cut it with my little hatchet. '' The anger died out of his father's face, and taking the boy tenderly in his arms, he said: "My son, that you should not be afraid to tell the truth is more to me than a thousand trees! Yes - though they were blossomed with silver and had leaves of the purest gold!''

legend George Washington Augustine Washington Mary Ball Washington When George was about six years old, he was made the wealthy master of a hatchet of which, like most little boys, he was extremely fond. He went about chopping everything that came his way. One day, as he wandered about the garden amusing himself by hacking his mother's pea sticks, he found a beautiful, young English cherry tree, of which his father was most proud. He tried the edge of his hatchet on the trunk of the tree and barked it so that it died. Some time after this, his father discovered what had happened to his favorite tree. He came into the house in great anger, and demanded to know who the mischievous person was who had cut away the bark. Nobody could tell him anything about it. Just then George, with his little hatchet, came into the room. "George, '' said his father, "do you know who has killed my beautiful little cherry tree yonder in the garden? I would not have taken five guineas for it!'' This was a hard question to answer, and for a moment George was staggered by it, but quickly recovering himself he cried: "I cannot tell a lie, father, you know I cannot tell a lie! I did cut it with my little hatchet. '' The anger died out of his father's face, and taking the boy tenderly in his arms, he said: "My son, that you should not be afraid to tell the truth is more to me than a thousand trees! Yes - though they were blossomed with silver and had leaves of the purest gold!'' honesty