Opinions on few recommender systems

Like it is said in the lecture notes, interpretation of the rating scale can vary much. That's one reason why I like to see comments next to rating. I like to look for reviews and ratings for example about restaurants and certain products and I find it more informative to read what people actually think about the food or the item, than just to see a "4 star" rating.

I have used product suggestions in finding the items that I'm interested in. For example Amazon's book suggestions have helped me in finding good books and reviews made buying decision easier. It's nice to have some peer support in making decisions and finding the items in online shops too :).

I find implicit preferences one thing that encourage to navigate in a webpage. For example if we look at Youtube, we can see the most popular clips on the first page. This popularity is calculated from the times the clip has been viewed (at least I think so, correct me if I'm wrong). So yes, those clips on the first page can be the most viewed but are the the most popular..? In my opinion, no. Like it was mentioned in the lecture notes, watching a part of a video doesn't tell wether you like it or not.


References
Lecture notes: Recommender system slides
(Original slides by Juha Leino)