1 MILLION dollars in this thread.. ( programmers )

[SmokeyTheBear]

www.netflix.com

netflix is looking for a better rating system for its online dvd store

contest starts today , prize is $1,000,000

To win the prize, a software program must improve the accuracy of Netflix's current movie recommendation system by at least 10 percent. The differences will be tracked by a program that quantifies how well the recommendation systems predict which movies will be liked or disliked by a profiled consumer.

If there is no winner, the company will pay $50,000 annually to a contest entrant who makes the most progress toward the final goal.

[Chris]

that is a nice chunk of fuckin change

i know HTML :-D

[polle54]

$1,000,000?

that's quite a load of money

I am half way a software engineer and I will definately give this a shot

where can I see the rating system as it is right now?

[Antonio]

kind sirs, we can help you.....

[Klen]

I can bet they did that for traffic.

[Ice]

Link to contest?

[Chris]

$1,000,000?

that's quite a load of money

I am half way a software engineer and I will definately give this a shot

where can I see the rating system as it is right now?

netflix.com

[FuqALot]

where's the press release or original page about this

[SmokeyTheBear]

that is a nice chunk of fuckin change

i know HTML :-D

you only really need the idea, if you have an idea you could split it with a programmer who would make it work

[Sosa]

wow netflix really throwing the cash out there

[SmokeyTheBear]

where's the press release or original page about this

whoops forgot to post that

http://news.yahoo.com/s/ap/20061002/.../netflix_prize

[OzMan]

kind sirs, we can help you.....

how's the weather in Mumbai at the moment?

[Libertine]

Hmmm...

An artificial neural network that learns to predict what movies people would like based on factors like genre, keywords and past ratings could work pretty well. The only problem is that it would take too long for the system to learn what works and what doesn't, even if you allow it to use data already available.

[SmokeyTheBear]

Hmmm...

An artificial neural network that learns to predict what movies people would like based on factors like genre, keywords and past ratings could work pretty well. The only problem is that it would take too long for the system to learn what works and what doesn't, even if you allow it to use data already available.

well i think the trick is , dont try to get the system to work the trick is to beat the software they are going to use to test it

Like how could they possibly predict if the software works without putting it into effect, they dont , they will use a simulation based on their own data

[SmokeyTheBear]

use the old spinaltap method , take 10 off the rating system so it only goes up to 9 and u will have improved it 10%

[FuqALot]

people from Quebec not welcome

[pornguy]

Smokey. You can do that.

[SmokeyTheBear]

people from Quebec not welcome

probably , quebec has fucked up contest laws

[SmokeyTheBear]

Smokey. You can do that.

maybe , doubtfull , i will surely put my noggin to some floggin thinking about it , but i know theres alot of better programmers out there in more need than i

[darksoul]

kewl, I'm gonna win this

[Damian_Maxcash]

They need to buy data from 3rd parties - ok - where do I collect the check?

[SmokeyTheBear]

They need to buy data from 3rd parties - ok - where do I collect the check?

good answer

[gornyhuy]

Thanks for the heads up... I'll invite you to the n3rd victory party. ;)

[LittleSassy]

that's a lot of money....

[RayBonga]

I'll just hire a team of Indian engineers for $10k and pocket the difference.

Thanks for the tip.

[Trafficbrokercom]

The Amazon recommendations are sick.

The last seven CDs and books they recommended to me where all CDs and books that I did already buy offline during the last weeks.

[RobbieRye]

So far, the best solution has come from a 15 year old kid.

Good idea too.

[cyberstar]

wow.. that's a lot of money.. i wonder how much they make that they can afford to give away 1 mil!

[After Shock Media]

Never even used it before so no clue as to what they do now.
However could they not just pre survey you when you sign up or whatever.
Like ask you to list your favorite genre, pick your favorite 10 movies of all time, your favorite actor, actress.

From there ask you to rate certain movies from never seen, to great movie. Just a small selection of course, like 10 movies that were based on your pre survey info.

After that I am pretty certain you could start giving recommendations, and then hopefully people would rate rented movies when they return them to better suit the individual.

[cyberstar]

The Amazon recommendations are sick.

The last seven CDs and books they recommended to me where all CDs and books that I did already buy offline during the last weeks.

actually their recommendations are so good that i saved one of the emails they sent me for later... that's pretty good marketing!

[woj]

well i think the trick is , dont try to get the system to work the trick is to beat the software they are going to use to test it

Like how could they possibly predict if the software works without putting it into effect, they dont , they will use a simulation based on their own data

yea, but wouldn't you have to beat all other entries as well? 10% I think is the minimum goal, if someone comes up with 80% improvement, and you only get 15%, you don't get shit I would assume...

[sicone]

They can hire me for 500k a year and I will randomly select movies I think people should see, much more effient then any program plus they save 500k

[SxDx]

great contest. wish that i had a clue about programming!

[Trafficbrokercom]

So far, the best solution has come from a 15 year old kid.

Good idea too.

where did you find that info ?

[psili]

Guess I won't be participating.

After reading the rules, I had to take a couple of aspirin to kill my headache.

[Trafficbrokercom]

Guess I won't be participating.

After reading the rules, I had to take a couple of aspirin to kill my headache.

Contest begins October 2, 2006 and continues through at least October 2, 2011.

:D

[L0rdJuni0r]

nice chunk of change.. i wish i knew programming.

[fris]

[BitAudioVideo]

use the old spinaltap method , take 10 off the rating system so it only goes up to 9 and u will have improved it 10%

hahah "but this one goes to 11"

[studio]

smokey are you going to give it a whirl?

[emthree]

Hold on while I get india on the phone.

[xlogger]

Its $1m because its fucking impossible to do it.

They must already have like 40 - 50 in-house programers trying to make the system better.

My friend works in a company where they hire just shit load of programers to make the software better. Everyone is working on so its like its max, its really have to improve it.

[madawgz]

wow, thats alota money for a small script...

[nolaman]

exxxcellllent

[Aquarius]

Wow, nice offer for sure.

[FightThisPatent]

cross reference to the kevin bacon six degrees of separation database with the leading star cross-referenced by genre, and taking the first result after some weighting based on other movies the person has ranked


Fight the solution!

[Elli]

All I know is Rogers Video recommendations suck monkey balls.

That would be one heck of a program. I bet the FBI has just what they're looking for.

[griffin]

Haha you guys acting like this is easy and a small script make me laugh. Netflix has a number of very smart programmers working on their system already. *Hint* anything a business is willing to pay $1million cash for isnt easy or they'd hire 1 java programmer in banglore to do it on the cheap.

[Damian_Maxcash]

Haha you guys acting like this is easy and a small script make me laugh. Netflix has a number of very smart programmers working on their system already. *Hint* anything a business is willing to pay $1million cash for isnt easy or they'd hire 1 java programmer in banglore to do it on the cheap.

Haha - you guys that think anyone in this thread is being serious make me laugh.

[squishypimp]

is this for real?