Ask HN: Can some one point me to starters read on Multi-armed Bandit problems?
I found some articles but they assumed that I have some background in the area. Can I get a starters guide quickly? I want to use it for news article recommendations.
1 comments
eh, not sure if it's a business application or what, but I've dealt with similar things in code. To me the best way to do it is through recursive callbacks and modeling the problem over a large data set. Recently I've gotten into shoveling the output into Eureqa and letting it come up with close-fitting (unreadable and basically nonsensical) equations to model the results for faster processing of additional input after the fact. If I understand you correctly. Sorry I can't help more, it's not my area of expertise exactly...just something I've (been forced to) dabble with...
hmm...similar situation. But mostly an academic scene. Have to perform some tests using MAB and I am asked to see if I can model my recommendation problem into MAB. But anyways thanks for the update. Do you have any quick pointers to start?