typodupeerror

## Close but no Cigar for Netflix Recommender System114

Ponca City, We Love You writes "In October 2006, Netflix, the online movie rental service, announced that it would award \$1 million to the first team to improve the accuracy of Netflix's movie recommendations by 10% based on personal preferences. Each contestant was given a set of data from which three million predictions were made about how certain users rated certain movies and Netflix compared that list with the actual ratings and generated a score for each team. More than 27,000 contestants from 161 countries submitted their entries and some got close, but not close enough. Today Netflix announced that it is awarding an annual progress prize of \$50,000 to a group of researchers at AT&T Labs, who improved the current recommendation system by 8.43 percent but the \$1 million grand prize is still up for grabs and a \$50,000 progress prize will be awarded every year until the 10 percent goal is met. As part of the rules of the competition, the team was required to disclose their solution publicly. (pdf)"
This discussion has been archived. No new comments can be posted.

## Close but no Cigar for Netflix Recommender System

• #### 1.57% (Score:2)

Is the new margin of improvement for victory then?
• #### Re: (Score:2)

I would think they would not implement the AT&T team's solution given it did not reach the 10% mark, however AT&T has the lead in reaching that mark unless someone comes up with some quantum leap in design./p.

• #### Re:1.57% (Score:5, Funny)

<bittercode@gmail> on Wednesday November 14, 2007 @11:37AM (#21349547) Homepage Journal
Right - but the AT&T method is public - so if you can get that the rest of the way to 10%..... I guess step 3 is profit.
• #### Re: (Score:2)

Actually, and my math will suck here. If they implement the new AT&T data, and then ask for 10% it would be much harder then what AT&T themselves did.

If they were at 50% accuracy and AT&T gave them 8.63%, if they implement that they are now at 58.63% accuracy. If they require a 10% increase then a new person will have to bring them up to 68.63% accuracy, much harder then the 60% AT&T was aiming for. Assuming that it becomes harder as you get closer to 100% accuracy.
• #### Re: (Score:2)

I gave AT&T an extra 0.2% because I love Big Brother. Not because my math sucks.

-Dick Cheney
• #### Re: (Score:2)

My guess is that they're giving the award for cutting out 10% of the inaccuracy, ie if they're at 50% and you can get them to 55%. At that point, another 10% would be 4.5% instead of 5%. This is because there's almost no chance of getting to 100% probability, so you're going to the limit of 100% without any chance of getting there.
• #### Not how it works (Score:5, Informative)

on Wednesday November 14, 2007 @11:42AM (#21349599)
That's not how the contest works. It's based on the RMSE that the original netflix algorithm got at the beginning of the contest. This is fixed and does not change. See the contest site for more details.
• #### diminishing returns (Score:3, Interesting)

From my experience with the Netflix Prize, and ML/stat.learning techniques in general, that last 1.57% is going to be the hardest. There is a diminishing returns effect going on here, i.e. the effort required for each successive 1% increase gets progressively larger.
• #### Don't meme me bro (Score:2)

Any chance of not tagging this story with this meme?
• #### I'd say... (Score:5, Insightful)

on Wednesday November 14, 2007 @11:05AM (#21349251) Journal
If the people who created Netflix's system are still with the company, I'd say they deserve some retroactive recognition (and bonuses). That's pretty damn good optimization if it's that hard to improve upon, and there seem to have been some really sophisticated people trying to beat them.
• #### Re:I'd say... (Score:4, Funny)

on Wednesday November 14, 2007 @11:11AM (#21349307)
Perhaps they should look at whatever chooses the slashdot page-bottom quotes for inspiration: Mosher's Law of Software Engineering: Don't worry if it doesn't work right. If everything did, you'd be out of a job.
• #### Re: (Score:3, Interesting)

It's hard to say. On the one hand, it could be that the current system is good enough that improvements are minutely incremental, though 8+% is pretty good if you ask me. On the other, it may be that the system is so fraught with dependencies and/or the relationships are so variable that it's hard to make gigantic leaps in sophistication. Look at Amazon's recommendation system: pretty good overall but still makes some egregious errors. Add the tagging system to the mix and it's possible to lead the recommen

• #### Re:I'd say... (Score:5, Funny)

on Wednesday November 14, 2007 @11:41AM (#21349579) Homepage Journal
Customers who bought the items in your shopping cart also bought:

Empress Charmeuse Silk Sheet Set - Queen - Ivory ~ \$399.00
Black Leather Victorian Vintage Shaper Corset Boned Lace Up corset ~ \$119.99
Orgazyme Clitoral Stimulation Gel, Topical, 0.8 oz ~ \$20.79
Pampers Cruisers, Size 4, Economy Plus Pack, 140 Cruisers ~ \$38.99

I don't think Amazon has much room to improve their recommendation technology.

• #### Re:I'd say... (Score:5, Informative)

<fairwater.gmail@com> on Wednesday November 14, 2007 @01:21PM (#21351167) Homepage

ook at Amazon's recommendation system: pretty good overall but still makes some egregious errors.

Egregious errors? It's downright useless unless you pretty much buy only one genre of book/music/whatever. Their system is heavily weighted towards whatever you most recently bought - and drops huge slabs of quasi related stuff into your request list at the slightest provocation.

I buy (among other things) serious works of culinary history, sociology, etc... Yet my reccomendation list is clogged with food porn (coffee table cookbooks) and the latest crap offerings from whichever TV chef is the current flavor of the moment. It also doesn't recognize the difference between editions - if you buy a hardback, it'll happily reccomend you buy the paperback. If you buy a frequently reprinted SF novel, it'll happily add each new printing/edition to your queue.
• #### That's funny... (Score:1)

When I think of "Food Porn", I certainly don't think of coffee table cookbooks...
• #### Re: (Score:2)

But that's what they are called in the foodie world - glossy volumes with lots and lots of photographs and food, wonderful dining spaces, regional scenery, etc... And usually waaay to expensive to take into the kitchen and actually work from. (And not aimed at cooks anyhow - but at foodies.)
• #### Re: (Score:3, Interesting)

Most everyone who tried was able to beat Netflix's existing system. I put together a little framework to help people get up and running faster. A lot of people seemed to be spending time just getting all the data into memory before they got to play with any algorithm ideas. I include a few algorithms including Simon Funk's which should be enough to get you started. http://www.icefox.net/programs/?program=NetflixRecommenderFramework [icefox.net]
• #### AT&T Labs? (Score:2)

At the risk of getting marked redundant: I totally agree, and I don't work for Netflix

This is a great contest, considering they have to publicly release the solution.

Although what is AT&T doing working on this problem?
• #### Re: (Score:2, Insightful)

by Anonymous Coward
AT&T Labs = bunch of people from former Bell Labs = welfare for AI researchers ;)
• #### Re: (Score:2)

It's too bad they didn't win the full prize, but at least they now have lucrative job opportunities in the booming spam industry. I keep getting spammed with Viagra ads and illegal replica watches, but I really want all of the hardcore pr0n spam.
• #### Re:I'd say... (Score:5, Interesting)

on Wednesday November 14, 2007 @11:36AM (#21349545) Homepage
I think the problem is that (and I may be wrong) any new system that researchers come up with isn't allowed to ask the user for more information. This would make if very hard for any system to be acurate if it is based soley on what dvds you rented and how you rated them.
If I liked Die Hard 4, for example, did I like it because of Bruce Willis, the "I'm a Mac" guy, the special effects, the plot, or some other reason that even I don't know?
Personally, I know that I have rated something like 900 movies on the netflix site and nearly all the recommendations are things I've no interest in or they simply say, "Sorry we have no recommendations for you at this time."
I would like to think that if they could ask me why I rated one movie a 4 and another a 1 then they might have more accurate recommendations. Even if they just had a drop down menu with something like, "I liked this movie because of a) the starts, b) the plot, c) the genre and so on" it would make recommendations a lot easier.
• #### Re: (Score:3, Interesting)

Thats why they need to get multiple data points to make a recommendation. If you rented a lot of the "I'm a Mac Guy" movies and rated them highly, then there is a bigger chance that that is the reason you liked that movie. If you refused to rent, or rated poorly the movie "Rugrats Gone Wild" then you probably aren't a rabid Bruce Willis fan etc. The entire goal of the project is to find films you like without you having to do a mini-review of every movie you have rented/saw.
• #### Re:I'd say... (Score:4, Informative)

on Wednesday November 14, 2007 @12:03PM (#21349911) Journal

Thats why they need to get multiple data points to make a recommendation. If you rented a lot of the "I'm a Mac Guy" movies and rated them highly, then there is a bigger chance that that is the reason you liked that movie. If you refused to rent, or rated poorly the movie "Rugrats Gone Wild" then you probably aren't a rabid Bruce Willis fan etc. The entire goal of the project is to find films you like without you having to do a mini-review of every movie you have rented/saw.
One of the common approaches to recommender systems is SVD, or Singular Value Decomposition [wikipedia.org]. SVD tries to isolate "features" in the training set that best represent a particular trait of the data and its value, such as the examples above. You may not have any idea what the feature actually represents, but that is fairly common in machine learning. It is an iterative process. Once you have defined one feature as well as you can, you move on to a new one. There are diminishing returns with this approach though, and identifying too many features can overspecialize your system and yield worse results. If your results are not good enough, you can try a different approach. Once you have tried several approaches that are almost good enough, you can try combining the different results to varying degrees to get a hopefully better result. That is what the leaders have done so far.
• #### Re: (Score:1)

I think this problem is solved through cluster analysis. The same way to tell if someone has multiple disjoint interests or a family is sharing an account. The goal is to find a "type" of thing you like and predict other things in that bin.
Your predictions will be more accurate if you don't try to match with other people who liked Die Hard 4 and Finding Nemo, unless they only rent movies that have CGI/special effects.
• #### Re: (Score:1)

You don't need to explicitly say this to get the data. Movies have a lot of information in them. You can tell which actors are in it, the plot, and the genre. If you make two selections and they both have Bruce Willis in it you can assume that you like Bruce Willis. You do not need to ask these questions of the user.
• #### Re: (Score:2, Insightful)

The idea is that with enough data, you could extract the "why" automatically. For example, if you rated all Arnold Scwarzenegger 5, then its probably because you like Arnold. If however you gave a rating of 1 to Kindergarden Cop, as well as "The Game Plan" and a bunch of similar movies, the system could also infer, that as much as you like Arnold, you don't like kids movies starring washed up "action" movie stars.

This is the whole idea behind the field of "machine learning": inferring causes/relationship/
• #### Re: (Score:1)

One additional confounding variable is that my wife and I both rate movies on the same Netflix account. She tends to like romantic comedies, me, not so much. But on Netflix, you would see both romantic comedies and thrillers rated highly, even though for an individual person that connection may not make sense. So you would also need to build into the model a way to identify cases like mine where the ratings are really the results of two people.
• #### Re: (Score:1)

You know what movies people like, and which ones they don't. Compare the movies together, and you can tell that the only reason someone likes these movies is because the have macintosh in them. Anyway, people HATE giving feedback. No one likes filling out a questionnaire. It takes way too long. You suggested a drop down. That only allows a user to pick a single reason. What if they like a movie for multiple reasons? What if they like the stars, but like the plot more? Shouldn't you capture that? I
• #### Re: (Score:1)

I'm probably wronger...

If you rated Die Hard 4 highly, and rated lowly some other movie featuring the "I'm a Mac" guy, but rated highly another featuring Bruce, would it be hard to figure that out?

You rate some movie well, it should recommend a similar movie, see how you rate that and keep building the data set.

It wouldn't even have to know anything about the movie in particular, just how other people rated the same movies.

Find somebody who rated movies rather similarly to how you have rated your movies

• #### Re: (Score:3, Informative)

From the Netflix Prize FAQ, they say how they currently do it:

"Straightforward statistical linear models with a lot of data conditioning."

The Netflix programmers shouldn't necessarily get special recognition for using least-squares modeling, but feel free to pass on your praise to Gauss, Legendre, Galton, and Fisher.

What's amazing is how hard it is to improve drastically on these 150-year-old statistical techniques.
• #### Re: (Score:2)

"Straightforward statistical linear models with a lot of data conditioning."

There's a whole lot of devil in those details, though.

• #### Re: (Score:2)

There's a whole lot of devil in those details, though.

Absolutely. With the size, sparsity, self-censoring and ordinal nature of their dataset, a 'straightforward statistical linear model' would not get very far.

• #### It's a scam. (Score:2)

Netflix's system is already 90.3% accurate!
• #### Re: (Score:1)

It seems to me that it must be hard to optimize because of the 5 star system.

90% of the movies I rate are either 3 or 4 stars. I already pre-filter so I don't watch movies that would get a 2 or 1 often and 5s are hard to fine. Trying to differentiate the emotions generating "meh" and "yeah" is gonna be tough. I don't know if most people rate similarly but I imagine they do.

A 10 star system would add more data points and might be better. But a simple system with multiple axis would be a lot better, I bet
• #### Re: (Score:1)

5 stars is certainly not enough for me either. Most movies end up with 4 stars for me, but I'd rather have the option of 6-8 stars, since there is certainly a lot of wiggle room in a 4.
• #### Re: (Score:1)

Actually, most movies should fall into the '3' category, with those you watched but didn't like as 2's, and those you couldn't watch as 1's (even some technical 2's might become 1's). Anything you have no interest in should be tagged as such. 4's are for those you really liked, and 5 for outstanding movies which in most cases you'd watch several times and possibly buy. (Sixth Sense was a 5 but is a one time only movie unless you're interested in dissecting it. There's too many movies in my queue for that;)
• #### This is why... (Score:2)

When I did use NetFlix, I spent a good amount of time flagging as many movies I did NOT want and would never, ever rent as those I did or would. The result was a pretty consistent selection that reasonably matched my taste.
• #### Moving target? (Score:5, Interesting)

on Wednesday November 14, 2007 @11:13AM (#21349319)
Will Netflix incorporate the near-winners' ideas into their current system? If so, won't future teams be aiming at a moving (improving) target? If not, won't current Netflix customers know that their recommendations could be better if Netflix just incorporated a now publicly-disclosed algorithm into their servers?
• #### Re: (Score:3, Interesting)

I had the same thought. And to what extent is the accuracy of a suggestion system important? Sometimes, throwing in a completely different suggestion might garner you a rental and possibly more rentals because you might like other movies of that type.
• #### Trust (Score:2)

Accuracy in this contest is defined as the user rating highly the movies that the system would suggest to them. The whole point of it is trust. If you're throwing out lots of suggestions that the user doesn't like just to try to find one they might like, you're destroying their trust in the system. They won't bother even reading the recommendations if they know they're filled with garbage.
• #### Re: (Score:2)

I would like a "rent random un-rated, un-rented movie" option. Whereby if you've already rated or rented a movie it will be excluded, but your queue gets some random movie put in.
-nB
• #### Re: (Score:2)

Unfortunately (for you), you are probably not very representative of netflix's customer base. It's all about cost/benefit. For THEM, not you. ;)
• #### Re: (Score:1)

Agreed, I like many movies across a spectrum of genres. Just because I liked Saw does not mean I want to see a bunch of bad horror movies more than a good comedy.
• #### Re: (Score:2)

Sometimes, throwing in a completely different suggestion might garner you a rental and possibly more rentals because you might like other movies of that type.

It's very, very important. If it isn't highly accurate, you're just going to completely ignore what it suggests, and get no benefits from it.

And your analogy is extremely flawed. If it's a movie you would like, then it SHOULD be recommended. That's what the system is there for. The odds that recommending a random movie to someone will inadvertently

• #### Re:Moving target? (Score:5, Informative)

on Wednesday November 14, 2007 @11:30AM (#21349469)
It's not a moving target. It's a very fixed number (RMSE = 0.8563) that the winning algorithm has to come up with. The netflix algorithm never gets re-run on the data for the prize.

Netflix is free to merge any improvements into their actual system in the meantime.
• #### Re: (Score:1)

No. The rules state: To qualify for the \$1,000,000 Grand Prize, the accuracy of your submitted predictions on the qualifying set must be at least 10% better than the accuracy Cinematch can achieve on the same training data set at the start of the Contest. The official contest site can be found on http://www.netflixprize.com/ [netflixprize.com]
• #### Bad title (Score:5, Funny)

on Wednesday November 14, 2007 @11:25AM (#21349433) Homepage
The prize was clearly a million dollars, not a cigar! I guess the editors don't even read the summary.
• #### Re: (Score:3, Informative)

I don't think I'm mistaken in saying that \$50,000 buys quite a few cigars too.

Explain.
• #### Re:Bad title (Score:5, Funny)

on Wednesday November 14, 2007 @12:22PM (#21350159)
Money can be exchanged for goods and services.
• #### Re: (Score:2)

Explain

Marijuana affects the memory.
• #### Re: (Score:2)

Woo-hoo!
• #### Re: (Score:2)

Perhaps the summary was not recommended to them, given their personal preferences.
• #### Re: (Score:1)

They didn't win a million dollars, or a cigar... so really it doesn't matter.
• #### no breaktrough - just blending (Score:5, Informative)

on Wednesday November 14, 2007 @11:34AM (#21349529)

Most noteworthy aspect of the winning entry is that their winning method works by combining 107 different types of prediction strategies.

They state that you can get pretty far by blending the 3-4 best strategies, but of course doing so would not have netted them the progress prize

It is kind of sad realization that there actually is no better method. Your best bet is to use brute force and attempt to find some weighting methodology that combines known methods. By the way this is a well known issue in protein structure prediction competitions, for many years now so called meta-servers (predictions work by merely combining other predictions) win all the time. The joke is that we now need meta-meta-servers, combine the results of combiners

Also a clarification on the progress prize: to get it you need to have at least 1% improvement over the previous result. Considering that there is only 1.57% to go there is room for only one more progress prize until it hits the Grand Prize (10% improvement over the original results).

• #### Re: (Score:1)

Also a clarification on the progress prize: to get it you need to have at least 1% improvement over the previous result. Considering that there is only 1.57% to go there is room for only one more progress prize until it hits the Grand Prize (10% improvement over the original results).

Where did you get that? The rules (http://www.netflixprize.com/ [netflixprize.com]) state:
To qualify for a year's \$50,000 Progress Prize the accuracy of any of your submitted predictions that year must be less than or equal to the accuracy value established by the judges the preceding year.
You just have to be better.

• #### Re: (Score:1)

from: http://www.netflixprize.com//community/viewtopic.php?id=799 [netflixprize.com] We have also updated the Prize leaderboard to reflect the award of the 2007 Progress Prize and have established the new accuracy requirement to qualify for the 2008 Progress Prize. Again, in accord with the Rules, the new Prize level reflects a 1% improvement over team KorBell's verified submission, requiring a 9.34% improvement over the original Cinematch accuracy level.
• #### hmmm .... (Score:5, Funny)

on Wednesday November 14, 2007 @11:36AM (#21349543)
if (\$director eq "Michael Bay") {
print "Not recommended";
}

That should improve the system by at least 20%
• #### like trying to win the lottery (Score:2, Insightful)

I'm skeptical about these sorts of prizes. The X prize, Top Coder, Clay Institute Millennium Prizes-- if those were the only reasons to do something, few would. Seems pretty risky to do a lot of work for what amounts to a lottery ticket. So, who got 2nd place, and how well did they do? 1 group wins a paltry \$50K and a little publicity and recognition, maybe even an endorsement or two, and the other 27000 plus get what? Nothing much. It's cool and fun to work on such problems, but people have bills to

• #### The thrill of victory (Score:2)

I can say I played with it because I found it fun. I'm a coder, it's what my brain is interested in. There have been contests for ages simply because human beings like to compete, even if second place gets nothing.

And FYI, netflix doesn't get any "ideas" from anyone but the winner. You only have to submit your code if you win.
• #### Re: (Score:2)

Well, I'll tell you something. Most criminals suck at judging risks. They simply tend to forget that robbing a bank is very likely to put them in jail. On the other hand, some people are afraid of speeding even though they're driving on a road that is extremely unlikely to be patrolled by cops. My point is that different people have completely views on risks and I think it's extremely rare that people actually back up their actions with maths and statistics, instead its all about emotions.

I think its good t
• #### Re: (Score:2)

Except that the only risk here is wasting your time...
• #### Re:like trying to win the lottery (Score:5, Insightful)

<foobarsoft@foobarsoft.com> on Wednesday November 14, 2007 @12:09PM (#21350003) Homepage

Two reasons I can think of. One is the challenge. I like to code but I'm not great with coming up with projects to do myself. This kind of thing would be nice for that.

The other is the experience. If you get second in this, no, you won't win the prize. But you can bet that having that on your resume would make getting many jobs much easier. Amazon would like your skills. So would many other retailers.

Also, as a side note, it's not a lottery. There is a three prong legal test in the US to determine if something is a lottery. I think the three parts are you have to pay to get it, everyone has an equal chance of winning, and there is no skill involved. I'm not positive about the second part. This is free to enter and is based quite a bit on skill, so it's not like a lottery.

Don't exaggerate.

This isn't a way to get free work. It's a way to get very smart job candidates to find you. It's a recruiting tool. You don't honestly think that they will take the winning idea, pay the \$1m, and then just say "bye" do you? They will offer that person a job if at all reasonable (if it's a team of 500 students, obviously they couldn't).

• #### Re: (Score:2)

The three legal tests are:

1) Prize
2) Chance
3) Consideration - you pay something to enter.

Eliminating item 3 is typically how sweepstakes are made legal.
• #### Re: (Score:2)

Haha you are thinking way to slow. Most everyone on the top list has already being contacted for jobs, not be Netflix, but all sorts of other companies.
• #### Re: (Score:2)

That's true, but the prize is just icing to most (all?) of these groups. Many will spend much more than the prize to get it, and everyone involved knows this and still goes through with it. Sometimes it's enough to do it to advance the technology. You can consider it a prize that millions of people will make use of what you produced. Also don't underestimate the fun factor. It's a big drive for what people do. It's a cliche but money isn't everything. Also, research groups could be working on this, while ge
• #### Re:like trying to win the lottery (Score:5, Insightful)

on Wednesday November 14, 2007 @01:04PM (#21350837)
"Any contestants reading this? Maybe you could enlighten the rest of us on why you bothered competing?"

There are two immediate reasons I can think of why anyone would bother competing:

1) To win money.
2) Because they enjoy the challenge of trying to solve an interesting problem.

I'm just a simple coder, and knew that I didn't have any realistic chance of winning money. But I still found it very satisfying to try to come up with a solution and send it in and see how I did. I don't regret spending hours of my own leisure time on the project.

That said, eventually I gave it up. It was very clear that I'm not smart enough to meet the challenge. I had my fun, and it was time to move on to the next project. In summary, I don't think it's safe to assume that everyone is in it for the money.

"Pardon my cynicism, but seems like contests like this are a way to get a lot of ideas and work for very little money."

I call it "brilliant". Netflix probably put some pricetag on what it would pay to get >10% improvement on their system. That pricetag is probably more than \$1 million. That means profit!
• #### Re: (Score:2)

There's a third reason: reputation. Being able to say "won the \$FOO Prize" is probably worth lots more in terms of future employment than the actual prize.
• #### partial credit? (Score:2)

They should give AT&T \$843,000.
• #### Man, that's too easy... (Score:2)

some

if age 18 and male then hard porn and south park
if age > 18 and male and lives at home then any sci-fi movie (plus points if it's a sequel)
if age > 18 and female then any movie with Princess in the title

100% match up !
• #### Re: (Score:2)

if age > 40 and female and single, suggest movie showing what pig men are but true love will eventually be found
if age > 40 and male and single, porn

that's what i get for listening to, uh, slashdot: http://slashdot.org/article.pl?sid=06/10/09/1344235 [slashdot.org]
• #### only yourself to blame (Score:2)

That's what you get for lack of reading comprehension, including familiarizing yourself with what the contest actually is. The contest wasn't to beat Netflix's algorithm. The contest was to beat it by 10%. Nothing in the summary of the original article was incorrect.
• #### Re: (Score:2)

surely i can be castigated for not having rtfa. however, given the fact the original article (http://developers.slashdot.org/article.pl?sid=06/10/02/1359221&tid=97) was seven days prior to the "beaten" story and neither of the writeups in any way mention the 10% threshold I dont feel like my confusion was particular unreasonable. asshole.
• #### Yes, please improve it. (Score:2)

I, for one, have really never found Netflix's recommendations all that useful. It sometimes recommends movies that I've already Netflixed. But to be fair I think they fixed that. It has recommended movies that I already have in my queue but most of the time it will be movies that I have no interest in at all. Then there was the time I turned in a 'G' rated movie, Disney I think, and it recommened ether Saw I or Saw II.

Not really sure where it got that one from. Nothing I had turned in that week had

• #### Throwing away too much information (Score:2)

They're looking in the wrong places, and trying to squeeze blood from a stone.

User ratings are a deeply flawed way of getting this information. They're one-dimensional and prone to serious randomizations based on the user's mood; a 5 today might have been a 3 tomorrow. Since most of the movies that a user rates will be between 3 and 5 (it's just not that hard to spot a movie you're going to hate, so why would you rent it in the first place?) that makes the data... well, not valueless, but containing a lot
• #### Sometimes throwing stuff away is good (Score:3, Interesting)

I think you have to consider that netflix is working off a very large user base with a very large list of titles. In this sense, computation time is going to go way up the more you keep adding all these factors. I'm sure they've had projects internal to netflix to use more data, but found that it just didn't pay off with the increased computation time. It's much better to get good recommendations onto the page instantly than make the user wait 2 seconds for great recommendations. The same is possibly tr
• #### Re: (Score:1)

You seem to have it figured out (even though you obviously haven't read the competetion rules and what information netflix uses for their current technique. So go give it a shot then, wiseass.
• #### Re: (Score:2)

What makes you think I haven't read the rules? I did read them, and downloaded the data, and decided not to participate because I didn't think that the existing algorithms could be usefully improved upon. The results seem to have borne me out.

Perhaps I should have phrased some of this better: what ELSE did they look at and decide against? What did they rent and change their minds about? (That's the second one and I really should have proof-read that better.)

I said in the original thread announcing the co
• #### They're not interested in "truth" necessarily (Score:2)

They just want to access the "truthiness" of recommendations :-)
• #### Re: (Score:1)

Let's look at that additional data a little closer.
Some of it might be useful, but a lot of it seems like noise.

* What did the user look at? - Noise, looking is way too weak an endorsement, and if not queued, it's probably not an endorsement at all

* What did the user rent? - Slightly more useful, but because the user wasn't explicit about their feelings (and most users don't rate a lot of movies), it's hard to come up with a convincingly reasonable way to rate them by default.

* How did they order their queu
• #### I rate 1 & 2 stars all the time (Score:2)

I've taken a bunch of Netflix recommendations. Some were brilliant and some were stupid, but I'm not afraid to list a movie as 1 or 2 stars, and I seldom rate a movie at 5 stars unless it was really, really good. I figure this helps the recommendation process.

I'd like to see 3 improvements, though:

1) Half-star ratings. I'm given recommendations in fractional star increments, and there are times where I think half-stars make sense -- there's been movies that haven't totally sucked, and 2.5 stars seems app
• #### Here's an idea... (Score:4, Insightful)

on Wednesday November 14, 2007 @04:18PM (#21353957) Homepage Journal
Why not give users more control over their recommendations? Heck, even a bunch of checkboxes would be useful.

For example, Netflix frequently recommends rated R movies to my family, but we have never rented a single R-rated movie and have no desire to do so. Moreover, every time we get a recommendation for an R-rated movie, we rate it "Not Interested." I've probably marked dozens of R-rated movies "Not Interested," but they continue to be recommended. (Either Netflix is trying to tell me to just give in and rent one already, or they really don't understand my family's movie preferences.)

A simple checkbox for "Do not recommend R-rated movies" would be all Netflix needs to substantially improve its accuracy for my family. I imagine Netflix could add checkboxes for similar criteria as well. In any case, I think a key point is giving more control over recommendations to the users themselves.
• #### Re: (Score:1)

I have no idea if this blocks suggestions, but I just went to netflix and each profile has a "maturity level restriction". Here's a tiny bit of the question/answer to one of the FAQs about it. It would *seem to me* that they would block suggestions if you set this. Also, note that another answer says that if a user tries to add a "too mature" movie to their queue, they are prompted for the account owner's password.

It seems worth setting this and try. It seems to me that they'd definitely limit what mov
• #### Buried gem (Score:3, Interesting)

<samwyse@gmaiLAPLACEl.com minus math_god> on Wednesday November 14, 2007 @04:20PM (#21353985) Homepage Journal
The most interesting part of the research paper was this: "More specifically, if movie i was rated x days later than movie j, we multiply their similarity by exp(-x/600). The denominator 600 (days) was determined by cross validation, and reflects the fact that after two years, similarity decays by approximately a factor of 3." Apparently Joe Average's tastes in movies slowly evolve over time, and something you liked three years ago may not be that attractive today.

This raises the question, should someone's age affect the denominator? People in or just out of college generally see their tastes evolve quickly, while people in retirement homes might take decades to get tired of something.

I also wonder if this decay factor applies to other fields. Not just books or music, but toothpaste or politicians. In the US, your representative is presumably re-elected before your opinion has time to change much; the president just as you're getting tired of him. It makes me wonder how Senators get re-elected at all.
• #### Theoretical Limit? (Score:1)

How do we know whether it is even possible, theoretically, to improve it by 10%?

1. inherent randomness in each individual's ratings

if you give me a list of movies today to rate them, then the same list a week later, I probably would give inconsistent scores. the more randomness in it, the less predictable it is. hack, Netflix could have deliberately introduced some randomeness in it so that nobody could ever get the prize.

2. sample size

imagine there is a underlying theoretical model that drives us to rate t
• #### So, there's money in movie ratings, eh? (Score:1)

So Netflix has money to spend on improving movie recommendations.

I'm a subscriber--why not give me a little of that cash?

I mean, if my opinion of a movie is this valuable, I expect to be compensated for participating in the system.

And that's why I never rate anything on the internet--they haven't made it worth my time.

Actually, what I'd prefer is if Netflix would give me a list of movies recommended by a group of professional reviewers I tend to agree with.

And at \$4.99/month, that list doesn't have to be mo
• #### netflix (Score:1)

You know, I have actually had some luck with the system. Some suggestions were flops, but some were good.

#### Related LinksTop of the: day, week, month.

If you fail to plan, plan to fail.

Working...