Surfing The Laws of Media

Friday, March 11, 2005

Developing the Ratings

The other title could be "Keeping the Kooks at Bay".

More than academics are interesting in Marshall McLuhan. Conspiracy theory crazies and who knows who may find the "answer" in McLuhan's poetic use of the language. Amazon, eBay and other community web site encounter this problem and use community ratings to solve it.

Surprisingly none of the community web books that I own even have "ratings" in their index. I was hoping to find an existing algorithm to help me build a rating system. It looks like I'll have to be creative until help arrives from the web or friends.

So here are my thoughts. First, every tetrad can be rated. Second, until otherwise determine, everyone's ratings are given equal weight. Third, everyone's ratings are not equal. This is determined by other members of the community. Fourth, a statistically significant sample may not be possible so an external factor may be needed. For example, if a college professor is a member, that might enhance the significance of their ratings. Fifth, the overall rating of any tetrad should be the sum of all ratings where each individual rating is multiplied by the rating of the person suppling the rating. OR do it like amazon with two different rating systems - one for comments and one casual.

Overall observation - ratings and comments are very closely tied. Maybe a comment service has this nailed.

I want to make this simple to start.  Very simple. Cannot afford the overhead of an like system.

So there you have it. The problem of the night.



Post a Comment

<< Home