Relevancy Algorithm

[yellow281]

Greetings all,

 

This is my first post on Hypography and I am more than certain that I am in the company of some extremely smart folks. With that said, I am working on a project that requires an algorithm capable of determining relevancy when delivering content to a recipient. Relevancy is displayed to the recipient as a score from 1 to 100. Many factors determine the level of relevancy to the recipient and while I know what these factors need to be, I am at a loss as to how to create and document such an algorithm let alone program it.

 

The project in question is a start up so while resources are limited, I'll do whatever I can in order to compensate someone that could help with my quest!

 

Many thanks in advance and wherever you are, happy Wednesday =)

 

Kate

[Turtle]

hi kate. welcome to the menagerie.

 

i'm not the programming master, but a couple things struck me while we wait for one. first, can you list the factors you know for us? the other bit is that besides getting relevancy data from the specific search terms the recipient uses, do you have a record for recipients past searches?

 

that's all i got.

[yellow281]

Absolutely! I guess there is no way to make this discussion private though huh?

[Turtle]

Absolutely! I guess there is no way to make this discussion private though huh?

 

we do have private member-to-member messaging, but in the past when i have asked the wizzards here for help in programming some of my number theory hunts, i found the best results by posting publically. several people checking for errors while trying to one up the other(s) is a developmental dynamic not found in the closet.

 

ps does that "absolutely" refer to listing some factors, or that you have a data set for each recipient, or both?

[Turtle]

afterthoughts:

 

i happened to catch the PBS NOVA series program Hunting the Hidden Dimension last night & i got to thinking that it may have some ideas for you. inasmuch as a measure of relevancy for you has elements of recursive patterning, maybe we could develop a fractal algorithm. is relevancy self-similar at different scales? do you think anyone has looked?

 

here's a link to that program: >> NOVA: Hunting the Hidden Dimension

[phillip1882]

i'm a half decent programmer so may be able to give soem advice. i would obviously need more info. what factors exactly determine relevancy, and how do they do so?