Electronic Commerce Research Laboratory 网络安全实 电子商务研究实验室网络安全实.pptVIP

Andare piu’ veloce qui * The Spearman footrule distance (of two ranked lists) is the sum over all the elements of the absolute difference of their rank. * Kemeny distance is the number of pairwise disagreements. So it is the number of permutations to modify one ranked list in the other. Both rank aggregations are optimal but they are different. Both aggregations are kemeny optimal. Because a better one should have 0 distance from one of the two lists (to get global distance =1) but the first one has distance 2 from the second list (and vice versa) so the minimum is 2. * * This function is maximized by any sorting of the test items in decreasing value of their ratings Andare veloce qui – dire solo quello di sopra. * * pseudo-random伪随机：group中用户随机聚合，但是内部成员之间存在一定的关系 * 为什么随机的方法nDCG值也很高，主要原因在于对于一个长度为5的排序序列，随机排序平均nDCG值都能达到0.91，所以会很高 * * C’e’ quella linea verticale sopra perche’ il massimo del gain e’ comunque quello per il group recommendation che ottiene NDCG = 1. Infatti, per esempio quando la individual ha NDCG = 0.95 il gain puo’ al massimo essere 0.05, ovvero quando la group recommendation vale 1. * C’e’ quella linea verticale sopra perche’ il massimo del gain e’ comunque quello per il group recommendation che ottiene NDCG = 1. Infatti, per esempio quando la individual ha NDCG = 0.95 il gain puo’ al massimo essere 0.05, ovvero quando la group recommendation vale 1. * This result motivated the following experiment * where we look at the correlation between the average similarity of a user to the other group members and the effectiveness of the group recommendations for this user. * Rank aggregation techniques 为group recommendation提供了实现的可能 * * Group Recommendations with Rank Aggregation and Collaborative Filtering Linas Baltrunas, Tadas Makcinskas, Francesco Ricci Free University of Bozen-Bolzano Italy fricci@unibz.it Motivations Rank aggregation techniques are useful for building meta search engines, selecting documents satisfying multiple criteria, and spam reducti