Kurs:Recommender System
| Lfd. | Titel | Abstract | Bewertung |
|---|---|---|---|
| Scalable Machine Learning (CS281B) Recommender Systems Part 1 | 492 views | ||
| Scalable Machine Learning (CS281B) Recommender Systems Part 2 | Ever wonder how netflix can predict what rating you would give to a movie? How do recommendation engines get built? Well, it's possible with JRuby and it's fairly straight forward. Many engines are built purely on support vector machine regressions which map arrays of data onto a classifier, like a star. In this talk I'll explain how support vector machines are built, and how do make a simple movie prediction model all in JRuby. | 344 views | |
| Recommendation Engines using Machine Learning, and JRuby by Matt | |||
| Deepak Agarwal: Recommender Systems - The Art and Science of Matching Items to Users | 2378 views | ||
| Recommender System based on Model Similarity at the Mining Software Archives (MSA) 2010 | Chit-Chat-Talk about Recommender System based on Model Similarity at Mining Software Archives (MSA) 2010, Ascona Switzerland | ||
| Recommender Problems for Web Applications, 4/26/2010 | Please also see http://www.sfbayacm.org/?p=1579
for links to the presentation slides and other related documents. DESCRIPTION: Several web applications like content optimization and online advertising involve recommending items from an inventory for each user visit to maximize some yield metric of interest (e.g. click rates). These are instances of large scale recommender system problems that entail several statistical challenges. We provide a mathematical description of the problem followed by modeling solutions for a content optimization problem that arises in the context of Yahoo! Front Page (www.yahoo.com). In fact, we discuss models to a) serve most popular items, b) serve items that are most popular in different user segments and c) provide personalized item recommendations for each user. Our models are based on time series methods, multi-armed bandit schemes and bilinear random effects model. One class of bilinear random effects model we propose extends reduced rank regression to incomplete matrices, the other class extends matrix factorization to incorporate covariates. Throughout, concepts are illustrated with examples and results obtained from "bucket tests" conducted on a real system. SPEAKER BIOGRAPHY: Deepak Agarwal is currently a Principal research scientist at Yahoo! Research. Prior to joining Yahoo!, he was a member of the statistics department at AT&T Research. He is a statistician interested in scalable modeling approaches for large scale applications. He has done extensive research on large scale hierarchical random effects model, computational advertising, modeling massive social networks with applications to call graph that arise in the telecommunications industry and modeling massive dyadic data that arise in applications like recommender systems. He has won four best paper awards (JSM 2001, SDM 2004, KDD 2007, ICDM 2009) that are directly related to the material of the talk. He has also done research in anomaly detection using a time series approach and computational approaches for scaling spatial scan statistic to large data sets. He regularly serves on program committees of data mining and machine learning conferences. He is currently associate editor for Journal of Americal Statistical Association, the top journal in the field of Statistics. He have given two tutorials on Statistical Challenges in Online Advertising at CIKM 2009 and KDD 2009. Deepak in collaboration with his co-authors have developed algorithms for real recommender systems that have been successfully deployed and thus has experience with both practical and scientific issues that arise in such applications. More info available at http://www.sfbayacm.org/?p=1579 |
||
| Building Web Reputation Systems | Google Tech Talks
July 1, 2010 ABSTRACT Presented by Randy Farmer.
What do Amazon's product reviews, eBay's feedback score system, Slashdot's Karma System, and Xbox Live's Achievements have in common? They're all examples of successful reputation systems that enable consumer websites to manage and present user contributions most effectively. Randy will talk about these examples and why reputation systems are critical for any organization that draws from or depends on user-generated content. "Randy" Farmer has been creating online community systems for over 30 years, and has co-invented many of the basic structures for both virtual worlds and social software. His accomplishments include numerous industry firsts (such as the first virtual world, the first avatars, and the first online marketplace). Randy worked as the community strategic analyst for Yahoo!, advising Yahoo properties on construction of their online communities. Randy was the principal designer of Yahoo's global reputation platform and the reputation models that were deployed on it. Slides available at http://www.slideshare.net/soldierant/5-reputation-missteps-and-how-to-avoid-them |
||
| Recommenders Everywhere: The WikiLens Community-Maintained Recommender System | Google Tech Talks
November, 13 2007 Suppose you have a passion for items of a certain type, and you wish to start a recommender system around those items. You want a system like Amazon or Epinions, but for cookie recipes, local theater, or microbrew beer. How can you set up your recommender system without assembling complicated algorithms, large software infrastructure, a large community of contributors, or even a full catalog of items? WikiLens is open source software that enables anyone, anywhere to start a community-maintained recommender around any type of item. We introduce five principles for community-maintained recommenders that address the two key issues: (1) community contribution of items and associated information; and (2) finding items of interest. Since all recommender communities start small, we look at feasibility and utility in the small world, one with few users, few items, few ratings. We describe the features of WikiLens, which are based on our principles, and give lessons learned from two years of experience running wikilens.org. Slides at http://www.cs.umn.edu/~dfrankow/files/wikilens12.ppt Speaker: Dan Frankowski Dan Frankowski is both computer science researcher and practitioner in software and algorithms development. He got his master's degree in computer science from the University of Minnesota in 1993, then spent a year in Budapest on a Fulbright grant studying mathematics. From 1997 to 2003 he was an algorithms guy at Net Perceptions. From 2003 to 2006 he was a research fellow with the GroupLens research group at the Unviersity of Minnesota, which is most well-known for recommenders, but now studies online community more broadly. He now works as a software engineer for Google Groups. |
8439 views | |
| Social Recommendations | Google Tech Talks
April, 10 2008 ABSTRACT Social Recommendations will change both the lens through which we see the world as well as the manner in which we experience it. Everything from the media that we consume to the events we attend will be influenced by hyper-relevant results delivered through hierarchical social relationships. This talk demonstrates current efforts to integrate social relationships into recommended user experience including SoMR, the Social Media Recommendation API. Speaker: Dan Carroll Dan is the Director of the SoMR (Social Media Recommendation) project and the CEO of imp, the Intelligent Media Platform. Dan has worked in magazine and book publishing, labor organizing, and at a public policy think tank. He holds a patent in digital media distribution and writes the blog www.mediapatron.com. Dan lives in Mountain View, California and serves on the boards of Echolocations and InRadio. |
6885 views | |
| Similarity Search: A Web Perspective | Google Tech Talks
October, 18 2007 ABSTRACT Similarity search is the problem of preprocessing a database of N objects in such a way that given a query object, one can effectively determine its nearest neighbors in database. "Geometric near-neighbor access tree" data structure, an early work (1995) by Sergey Brin, is one of the most known solutions to this problem. Similarity search is closely connected to many algorithmic problems in the web. Similarity search is an abstraction of many algorithmic problems we face in data management. In this talk we will focus on: - Personalized news aggregation: Searching for news articles that are most similar to the user's profile of interests - Behavioral targeting: Searching for the most relevant advertisement for displaying to a given user. - Social network analysis: Suggesting new friends. - Computing co-occurrence similarities. - "Best match search": Searching resumes, jobs, BF/GF, cars, apartments. We describe features that make web applications somewhat different from previously studied models. Thus we re-examine the formalization and the classical algorithms for similarity search. This leads us to new algorithms (we present two of them) and numerous open problems in the field. Speaker: Yury Lifshits Yury Lifshits obtained his PhD degree from Steklov Institute of Mathematics at S... |
17136 views | |
| PeB -Recommender Systems |