978 - 1 - 107 - 03607 - 9 - Statistical Methods for Recommender Systems Deepak
نویسنده
چکیده
A/A test, 67. See also online bucket tests A/B test, 8, 55. See also online bucket tests accuracy metrics, 59–62, 63 adaptive sequential design. See multiarmed bandit (MAB) problem Adomavicius, Gediminas, 261 Agarwal, Deepak K., 97, 99, 107, 110, 147, 198, 210, 242, 250t, 251, 252, 253, 261 algorithmic techniques, for recommender system, 5–7 alternating least squares, 35 AOL, 82t application settings, 85–87 context-related characteristics, 87 item-related characteristics, 85 response-related characteristics, 87 user-related characteristics, 85–86 architecture. See system architecture area under ROC curve (AUC), 62 ARS-based EM for logistic response, 157–161 E-step, 157–160 centering, 160 initial points for ARS, 160 M-step, 160–161 article recommendation, 9, 24, 85, 95, 134, 177, 186. See also Yahoo! Today module; Yahoo! Buzz; news article recommendation, multifaceted AUC (area under ROC curve), 62 Auer, P., 46–47, 108, 109 bag-of-words, 5 item characterization, 18–21 dense format, 19f dimension reduction, 19–20 phrases and entities, 19 sparse format, 18–19f term frequency-inverse document frequency version, 18 term frequency version, 18 unweighted version, 18 bandit scheme, 42. See also multiarmed bandit (MAB) problem Basu, Sugato, 261 Bayardo, Roberto J., 261 Bayes 2 × 2 scheme, 107 BayesGeneral scheme, 106–107, 112t Bayesian approach, to multiarmed bandit (MAB) problem, 42–46 Bayesian solution, most-popular recommendation, 98–107 BayesGeneral scheme, 106–107, 112 Gamma-Poisson (GP) model, 99 general solution, 104–107 dynamic set of candidate items, 105–107 nonstationary CTR, 107 two-stage approximation, 105 K × 2 case, 102–104 convexity, 104 Lagrange relaxation, 103–104 near-optimal solution, 104 separability, 104 one-step look-ahead, 99
منابع مشابه
Effect of Rating Time for Cold Start Problem in Collaborative Filtering
Cold start is one of the main challenges in recommender systems. Solving sparsechallenge of cold start users is hard. More cold start users and items are new. Sine many general methods for recommender systems has over fittingon cold start users and items, so recommendation to new users and items is important and hard duty. In this work to overcome sparse problem, we present a new method for rec...
متن کاملEvaluation of recommender systems: A multi-criteria decision making approach
The evaluation and selection of recommender systems is a difficult decision making process. This difficulty is partially due to the large diversity of published evaluation criteria in addition to lack of standardized methods of evaluation. As such, a systematic methodology is needed that explicitly considers multiple, possibly conflicting metrics and assists decision makers to evaluate and find...
متن کاملA Review of Spatial Factor Modeling Techniques in Recommending Point of Interest Using Location-based Social Network Information
The rapid growth of mobile phone technology and its combination with various technologies like GPS has added location context to social networks and has led to the formation of location-based social networks. In social networking sites, recommender systems are used to recommend points of interest (POIs) to users. Traditional recommender systems, such as film and book recommendations, have a lon...
متن کاملContext-Aware Recommender Systems: A Review of the Structure Research
Recommender systems are a branch of retrieval systems and information matching, which through identifying the interests and requires of the user, help the users achieve the desired information or service through a massive selection of choices. In recent years, the recommender systems apply describing information in the terms of the user, such as location, time, and task, in order to produce re...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
عنوان ژورنال:
دوره شماره
صفحات -
تاریخ انتشار 2016