« Sam's Story: Week 2 |Main| Book Review: When You Are Engulfed in Flames »
The computer says you like Police Academy movies February 9, 2004
Ever have one of those "Hey! They stole my idea!" moments? A while back my co-worker and I were discussing how statisticians could apply their powers to the betterment of mankind by helping them select movies, books, or video games based on self-reported ratings and data clustering methodologies.
The idea is pretty simple. First, you rate a bunch of books that you've read. Then a team of either trained chimps or graduate students make independent ratings about the books on a number of objective dimensions --things like publication date, genre, author, reading level, etc.
There's then a number of clustering methodologies that could be used to group books together based on these factors, putting them into "families" that may or may not transcend the traditional genre boundaries. Then, based on your ratings of the books you've read and those books' scores from the chimps/students, the system could assign weights to each factor and recommend new books to you. It could even look at people who gave similar ratings to the same books and use that information to further refine its recommendations. (In fact, this is probably similar to how Amazon.com's recommendations work.)
You could then really take it to the next level and further improve recommendations by factoring in demographic data (gender, age, education) or psychological measures like intelligence or personality. You could find that people who are high in emotional stability but low in conscientiousness tend to like mystery novels involving cats. Neat, eh?
It's kind of a one day project, in that we thought we could get to it "one day". Looks like someone beat me to it, though, because I was poking around the 'net today and noticed a site called What to Rent.
Their system purports to recommend movies based on your personality. When you sign up for an account you take a short "personality test" that asks you about your movie preferences. It then recommends movies and invites you to come back 24 hours later to tell it how you liked its recommendations by answering more questions. Over time it's supposed to get smarter and tailor its recommendations more snugly.
![]() |
They don't go into any detail about the methods used, saying that "The specific theory behind the computer system that recommends movie rentals has patents pending in many countries and cannot be published at this time due to concerns over uncredited use of the technology involved." I can tell you, however, that the personality "test" they give you is crap. It's just a short questionnaire about movie preferences, not a personality inventory based on a scientifically researched model. So that part of it seems like a sham --a glossy coat made to make a simple survey look more complicated than it is. This opinion is cemented by this quote:
To give a more specific idea how a personality is simulated and analyzed, displayed below is a schematic for the fear simulator that assigns virtual component values based upon the film database. While the actual implementation of the simulator uses Integrated Circuit chips, the design was done with analog components.
This circuit is operated with specified values and the generated output is graphed and evaluated at points of interest.That just makes ...no sense. That's an electrical diagram. I don't think they're even trying, and haven't completely discounted the possibility that the whole site is a scam or a hoax. Still, the vision is right, even if it's not there yet.
Comments (0) |
Permalink |
File under: I/O Psychology
all this copyright jamie madigan until the sun explodes | about this siteBy index
- All entries
- Sam & Mandy Stories
- Photos
By category
- Books
- Conversations
- Gaming
- General
- I/O Psychology
- Parenting
- Photography
- Samantha
- Site Maintenance
By tags
- Alphabetical Tag List
- Tag Cloud
By month
- November 2008
- October 2008
- September 2008
- August 2008
- July 2008
- June 2008
- May 2008
- April 2008
- March 2008
- February 2008
- January 2008
- December 2007
- November 2007
- October 2007
- September 2007
- August 2007
- July 2007
- June 2007
- May 2007
- April 2007
- March 2007
- February 2007
- January 2007
- December 2006
- November 2006
- October 2006
- September 2006
- August 2006
- July 2006
- June 2006
- May 2006
- April 2006
- March 2006
- February 2006
- January 2006
- December 2005
- November 2005
- October 2005
- September 2005
- August 2005
- July 2005
- June 2005
- May 2005
- April 2005
- March 2005
- February 2005
- January 2005
- December 2004
- November 2004
- October 2004
- September 2004
- August 2004
- July 2004
- June 2004
- May 2004
- April 2004
- March 2004
- February 2004
- January 2004
- December 2003
- November 2003
- October 2003
- September 2003




File under: