Technically, you should have a list of saved orders associated with a
customer ID. You should be able to search the saved orders for all customer
IDs that purchased the current product ID. Then, generate a giant list of
all products that all those customer IDs purchased, and compile them into a
big list, with the most common purchases at the top. You should then maybe
have an array looking like:

$Products["Product Name"] = Number Purchased;
---------------------------------------------
$Products["PHP Programming Book"] = 24;
$Products["MySQL Programming Book"] = 22;
$Products["PC Magazine Subscription"] = 14;
$Products["Windows XP Pro"] = 6;
$Products["Anti-Fungal Loin Cream"] = 1;
---------------------------------------------

Then, display any products that have more than X number of purchases. That
way, you truly have a valid "Customers That Bought This Also Bought These"
list. A bit intensive, maybe, but it works.

- Jonathan


-----Original Message-----
From: olinux [mailto:[EMAIL PROTECTED]]
Sent: Monday, January 21, 2002 3:25 PM
To: [EMAIL PROTECTED]
Subject: [PHP-DB] Similar Articles or Customers also bought...


I am trying to figure a simple way of creating a
script for a news site to recommend articles that are
similar [i.e. about the same company]

Two examples are:
http://www.theroyalgazette.com/apps/pbcs.dll/artikkel?Avis=RG&Dato=20011218&;
Kategori=BUSINESS&Lopenr=112180020&Ref=AR
  (See the "Linked Articles")

and Amazon's "Customers who bought this book also
bought:" feature

I think it would work something like this 

TABLE 1
keyword_id | keywords

TABLE 2
url_id | url | url_title

TABLE 3 
id | keyword_id | url_id

this would be your basic search engine - for example
you could look up a keyword in table 1 using that
keyword_id get the url_id's that correspond and then
get the url's and titles for display.

I guess that the "similar srticles" feature would work
the same way by taking the current url_id lookup the
corresponding keyword_id's use those to query for
url_id's that match those keywords and then grab the
url and title for display.

Confused? I am :)

This seems like it would work, but seems too
intensive... Maybe a different table structure which
contains url_id | similar_url would be better?
then you could look up the current url_id for url_id's
of similar articles?

Thanks for your help.

olinux



__________________________________________________
Do You Yahoo!?
Send FREE video emails in Yahoo! Mail!
http://promo.yahoo.com/videomail/

-- 
PHP Database Mailing List (http://www.php.net/)
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
To contact the list administrators, e-mail: [EMAIL PROTECTED]

-- 
PHP Database Mailing List (http://www.php.net/)
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
To contact the list administrators, e-mail: [EMAIL PROTECTED]

Reply via email to