Presentation is loading. Please wait.

Presentation is loading. Please wait.

Sarah Fatima Varda Sarfraz.  What is Recommendation systems?  Three recommendation approaches  Content-based  Collaborative  Hybrid approach  Conclusions.

Similar presentations


Presentation on theme: "Sarah Fatima Varda Sarfraz.  What is Recommendation systems?  Three recommendation approaches  Content-based  Collaborative  Hybrid approach  Conclusions."— Presentation transcript:

1 Sarah Fatima Varda Sarfraz

2  What is Recommendation systems?  Three recommendation approaches  Content-based  Collaborative  Hybrid approach  Conclusions 2

3 3 Items SearchRecommendations Products, web sites, blogs, news items, … Recommendation systems are programs which attempt to predict items that a user may be interested in

4  Main idea: recommend items to customer C similar to previous items rated highly by C  Movie recommendations  recommend movies with same actor(s), director, genre, …  Websites, blogs, news  recommend other sites with “similar” content 4

5  Consider user c  Find set D of other users whose ratings are “similar” to c’s ratings  Estimate user’s ratings based on ratings of users in D 5 Set of other users Similar Ratings Estimate

6  Works for any kind of item  No feature selection needed  New user problem  The same problem as with content-based system  New item problem 6

7  Implement two separate recommenders and combine their predictions  Add content-based methods to collaborative approach 7

8  Content-based  The user will be recommended items similar to the ones the user preferred in the past  Collaborative  The user will be recommended items that people with similar tastes and preferences liked in the past;  Hybrid  Combine collaborative and content-based methods 8

9  Explicit data collection  Ask people to rate items  Doesn’t work well in practice – people can’t be bothered  Implicit data collection  Learn ratings from user actions  e.g., purchase implies high rating  What about low ratings? 9

10  In most recommendation systems, consumers are asked to rate their liking of an item that they have experienced.  This is often done after receiving a system rating.  The accuracy of the recommender system is then measured by a mathematical comparison between the suggested system) ratings and the users’ actual ratings. 10

11  providing consumers with a prior rating generated by the recommendation system significantly influences their own rating in a way that artificially improves the resulting accuracy.  Experiment of a series of TV show conducted on 321 subjects. 11

12  due to its broad reach and open communication style, Twitter is able to facilitate intercultural communication in new and innovative ways.  In addition to reducing communication barriers, such as accents, text-based applications like Twitter chat provide a sense of universality to users. Likewise, by following others, users are able to develop a diverse social map.

13  Facebook reports “more than 750 million active users,” LinkedIn claims over 100 million members with 56 percent of those residing outside the United States and YouTube is “localized in 25 countries” and is available in 43 languages.  Employee recruitment and retention, internal communications, customer satisfaction and sales outreach are just a few of these activities taking place in this rapidly growing arena.

14  Crowd Sourced Wisdom  We share our knowledge, wisdom and experiences. Quora is a place where you can ask any question in the world, and expect a reasonable answer.Quora  Creativity and Inspiration is Unleashed  We share our creative ideas and inspirations. Pinterest has redefined the digital portfolio/catalog.Pinterest

15  Learn More from Each Other: We learn from each other. You can find everything from make-up tips to channeled extraterrestrial messages on YouTube.  Changed by Exposure to Diversity: We pay attention to the small details of people’s lives, delight over their family photos, and share at the level of family with a whole bunch of people. We’re supported when we need to be. We talk about issues and conundrums and joys. People are connecting and being exposed to both comfortable and diverse perspectives.


Download ppt "Sarah Fatima Varda Sarfraz.  What is Recommendation systems?  Three recommendation approaches  Content-based  Collaborative  Hybrid approach  Conclusions."

Similar presentations


Ads by Google