Anchoring effect in collaborative filtering

Wired has an article on Potter, the new Netflix competitor who took everyone by surprise — making it at the top of the list all of a sudden, passing off people who worked much longer in the competition. His insight is to correct for anchoring: a user who has recently given a lot of above-average ratings is likely to continue.

I worked on a similar effect back in 2003. Here is what I concluded:

We show that by normalizing users with respect to the mean, the amplitude, and, possibly, the number of their ratings, we improve accuracy. We stress that the normalization is per user as opposed to per item. (Scale And Translation Invariant Collaborative Filtering Systems, 2005)

Source: Andre Vellino.

Published by

Daniel Lemire

A computer science professor at the University of Quebec (TELUQ).

Leave a Reply

Your email address will not be published. Required fields are marked *

To create code blocks or other preformatted text, indent by four spaces:

    This will be displayed in a monospaced font. The first four 
    spaces will be stripped off, but all other whitespace
    will be preserved.
    Markdown is turned off in code blocks:
     [This is not a link](

To create not a block, but an inline code span, use backticks:

Here is some inline `code`.

For more help see