developing inteligently advertising shopping application

[nadeemshafi9]

hi guys

my application records what a user has looked at and how deep it has been looked at eg. looking at item = level 1 and more details level 2. it gets recorded like this :

usr_name  item_id    interest    time 
  user1      30            1        2007-01-15 14:57:06

ok i know my sql and php but have never actualy needed to use the more advanced SQL GROUP BY and sub queries etc. What i am trying to achive is and array of items that are similar to the ones i have analised as being the ones that the user is interested in most then intermediatly and then least,

but lets forget about that and get back to the begining.

my goal by posting this is to try and find some help in to the type of stuff i should be recording when a user browses an item and the type of analasys i should be doing on the data based on (your experiences as experinced developers), should i be trying to extract some sort of fuzzy logic and how because probability scales need to be based on 0 - 1, so the share of probability needs to be rippled throgh the whole table and thats quite hard.

is there any websites that you have come accross that talk about this stuff

thanks for any sugestions please feel free to comment in any way