18,824 Hits in 3.3 sec

Modeling and recovering non-transitive pairwise comparison matrices

Dehui Yang, Michael B. Wakin
2015 2015 International Conference on Sampling Theory and Applications (SampTA)  
-Arrow's impossibility theorem Form ranked list based on pairwise comparisons Pairwise Comparison Matrices • Let Y denote an n × n matrix where Y(i,j) represents the strength of preference of item i  ...  Rank Aggregation • Goal is to produce a single ranked list of n items (or candidates, teams, etc.) that best reflects the collective preferences of multiple voters. • Classical problem well studied in  ... 
doi:10.1109/sampta.2015.7148846 fatcat:k2o2uycjbzbsxjuaf2hjceef5e

IteRank: An iterative network-oriented approach to neighbor-based collaborative ranking [article]

Bita Shams, Saman Haratizadeh
2018 arXiv   pre-print
Finally, they use estimated pairwise preferences to infer the total ranking of items for the target user.  ...  of pairwise preferences to the target user based on the calculated similarities.  ...  Basic Mathematical Notations Notations Definitions U The set of users I Set of items P Set of pairwise preferences =< , > A pairwise preference of item j over item j C A three dimensional tensor  ... 
arXiv:1811.01345v1 fatcat:hajb7zitzjawjbdvjldiasymky

Initial Profile Generation in Recommender Systems Using Pairwise Comparison

Lior Rokach, Slava Kisilevich
2012 IEEE Transactions on Systems Man and Cybernetics Part C (Applications and Reviews)  
We present a new, anytime preferences elicitation method that uses the idea of pairwise comparison between items.  ...  Several systems try to learn the new users' profiles as part of the sign up process by asking them to provide feedback regarding several items.  ...  Here the value 1 indicates that both of the items are equally preferred. The value 2 shows that item A is slightly preferred over item B, etc.  ... 
doi:10.1109/tsmcc.2012.2197679 fatcat:opmdsf3wlnco7nsebjs5un2dqu

Neural Collaborative Ranking [article]

Bo Song, Xin Yang, Yi Cao, Congfu Xu
2018 arXiv   pre-print
We resort to a neural network architecture to model a user's pairwise preference between items, with the belief that neural network will effectively capture the latent structure of latent factors.  ...  In this paper, we examine an alternative approach which does not assume that the non-interacted items are necessarily negative, just that they are less preferred than interacted items.  ...  Weike Pan in Shenzhen university for helpful discussions on pairwise ranking techniques.  ... 
arXiv:1808.04957v1 fatcat:f7rn4jtn3rbzlmkka5ixamntbi

Pairwise Preference Regression on Movie Recommendation System

Rita Rismala, Rudy Prabowo, Agung Toto Wibowo
2019 Indonesian Journal on Computing  
Pairwise preference regression is a method that directly deals with cold-start problem.  ...  Recommendation System is able to help users to choose items, including movies, that match their interests. One of the problems faced by recommendation system is cold-start problem.  ...  Pairwise preference regression is one of the methods that directly deal with cold-start problem.  ... 
doi:10.21108/indojc.2019.4.1.255 fatcat:h34yh5ow2bhglggvmm3rjailqe

Adapting vector space model to ranking-based collaborative filtering

Shuaiqiang Wang, Jiankai Sun, Byron J. Gao, Jun Ma
2012 Proceedings of the 21st ACM international conference on Information and knowledge management - CIKM '12  
In this study, we seek accuracy improvement of ranking-based CF through adaptation of the vector space model, where we consider each user as a document and her pairwise relative preferences as terms.  ...  Experiments on benchmarks in comparison with the state-of-the-art methods demonstrate the promise of our approach.  ...  Based on the predicted pairwise preferences, a total ranking of items for user u can be obtained by applying a preference aggregation algorithm. Vector space model.  ... 
doi:10.1145/2396761.2398458 dblp:conf/cikm/WangSGM12 fatcat:tcnhgldkdbg5nchzmyksjx2f74

Improving Latent Factor Models via Personalized Feature Projection for One Class Recommendation

Tong Zhao, Julian McAuley, Irwin King
2015 Proceedings of the 24th ACM International on Conference on Information and Knowledge Management - CIKM '15  
Thus it may be better to view each dimension of a user's preference as a personalized projection of an item's properties so that the preference model can capture complex relationships between items' properties  ...  This matrix describes a mapping between items' factors and users' preferences in order to build personalized preference models for each user and item.  ...  Based on items' latent feature vectors and users' projection matrices, our PFP method models users' preferences over items in terms of projected latent feature vectors instead of a real number.  ... 
doi:10.1145/2806416.2806511 dblp:conf/cikm/ZhaoMK15 fatcat:iciw2kb2evfs3mlspfcssy3wpm

Clustering and Inference From Pairwise Comparisons [article]

Rui Wu, Jiaming Xu, R. Srikant, Laurent Massoulié, Marc Lelarge, Bruce Hajek
2015 arXiv   pre-print
In particular, we assume that there are $n$ users of $r$ types; users of the same type provide similar pairwise comparisons for $m$ items according to the Bradley-Terry model.  ...  Given a set of pairwise comparisons, the classical ranking problem computes a single ranking that best represents the preferences of all users.  ...  A net-win vector for a user is a vector of length m, where its i-th coordinate counts the number of times item i is preferred over other items minus the number of times other items are preferred over item  ... 
arXiv:1502.04631v2 fatcat:gl23lvhyybbyngqk5ssma6myhi


Stan Lipovetsky
2015 International Journal of the Analytic Hierarchy Process  
The results of stochastic modeling correspond to robust estimations of priority vectors not prone to influence of possible errors among the elements of a pairwise comparison matrix.  ...  The AHP priority vector can be interpreted as these probabilities belonging to the discrete states corresponding to the compared items.  ...  The preference of an i-th item over a j-th item corresponds to transition between them with intensity bij.  ... 
doi:10.13033/ijahp.v7i2.243 fatcat:dtkjhygmujg4phka3slrrwbv6i

Model-Based Learning from Preference Data

Qinghua Liu, Marta Crispino, Ida Scheel, Valeria Vitelli, Arnoldo Frigessi
2018 Annual Review of Statistics and Its Application  
items, when the user indicates only incomplete preferences; this is an important part of recommender systems.  ...  Preference data occurs when assessors express comparative opinions about a set of items, by rating, ranking, pair comparing, liking or clicking.  ...  For an unordered pair of items {A i , A k }, we denote a pairwise preference between the two items as (A i ≺ A k ), if item A i is preferred to item A k .  ... 
doi:10.1146/annurev-statistics-031017-100213 fatcat:vtykf5bp5zconbvy6mvwktqtby

General factorization framework for context-aware recommendations

Balázs Hidasi, Domonkos Tikk
2015 Data mining and knowledge discovery  
As context dimensions are introduced beyond users and items, the space of possible preference models and the importance of proper modeling largely increases.  ...  While these algorithms apply various loss functions and optimization strategies, the preference modeling under context is less explored due to the lack of tools allowing for easy experimentation with various  ...  A feature vector is defined for each property (including the item and the user itself), and the feature vector of the item (or user) is the weighted sum of the feature vectors of its properties.  ... 
doi:10.1007/s10618-015-0417-y fatcat:jc65vlzspbcfziyd6o2l5prvmu

DELF: A Dual-Embedding based Deep Latent Factor Model for Recommendation

Weiyu Cheng, Yanyan Shen, Yanmin Zhu, Linpeng Huang
2018 Proceedings of the Twenty-Seventh International Joint Conference on Artificial Intelligence  
A novel attempt of DELF is to model each user-item interaction with four deep representations that are subtly fused for preference prediction.  ...  Among various recommendation methods, latent factor models are usually considered to be state-of-the-art techniques, which aim to learn user and item embeddings for predicting user-item preferences.  ...  Latent factor models typically characterize users and items with feature vectors in the same latent space, and estimate each user-item preference based on the corresponding vectors.  ... 
doi:10.24963/ijcai.2018/462 dblp:conf/ijcai/ChengSZH18 fatcat:ve72jnzdrvc33a235zsecg75um

Ranking with Features: Algorithm and A Graph Theoretic Analysis [article]

Aadirupa Saha, Arun Rajkumar
2021 arXiv   pre-print
We consider the problem of ranking a set of items from pairwise comparisons in the presence of features associated with the items.  ...  \alpha\log \alpha)$, where $\alpha$ denotes the number of 'independent items' of the set, in general $\alpha << n$.  ...  From the user preferences, we first compute the underlying pairwise preference matrix P * , where P * ij is computed by taking the empirical average of number of times an item i is preferred over item  ... 
arXiv:1808.03857v2 fatcat:n5nojbwyprgrljzxcfto57mkca

Scalable Bayesian Preference Learning for Crowds [article]

Edwin Simpson, Iryna Gurevych
2019 arXiv   pre-print
We propose a scalable Bayesian preference learning method for jointly predicting the preferences of individuals as well as the consensus of a crowd from pairwise labels.  ...  As previous solutions based on Gaussian processes do not scale to large numbers of users, items or pairwise labels, we propose a stochastic variational inference approach that limits computational and  ...  Acknowledgements This work was supported by the German Federal Ministry of Education and Research (BMBF) under promotional references 01UG1416B (CEDIFOR), by the German Research Foundation through the  ... 
arXiv:1912.01987v2 fatcat:wpfk75m265atraoicmfbrlcnce

Item group based pairwise preference learning for personalized ranking

Shuang Qiu, Jian Cheng, Ting Yuan, Cong Leng, Hanqing Lu
2014 Proceedings of the 37th international ACM SIGIR conference on Research & development in information retrieval - SIGIR '14  
We propose a novel PRIGP(Personalized Ranking with Item Group based Pairwise preference learning) algorithm to integrate item based pairwise preference and item group based pairwise preference into the  ...  In this paper, we exploit this prior information of a user's preference from the nearest neighbor set by the neighbors' implicit feedbacks, which can split items into different item groups with specific  ...  Item Based Pairwise Preference The basic assumption of pairwise preference of two items can be formally represented aŝ rui >ruj, i ∈ I + u , j ∈ I\I + u whererui denotes preference of a user u on an item  ... 
doi:10.1145/2600428.2609549 dblp:conf/sigir/QiuCYLL14 fatcat:su36ib7a4bhivhw3mya7i55m34
« Previous Showing results 1 — 15 out of 18,824 results