Creating a "tag" system to find relevant "related" pages

MrQuestions · March 19th, 2008, 08:16 PM

Hello,

I'm wondering what everyone's opinion is on what would be the best way to implement comma-separated tags on a web site using PHP, MySQL, and Javascript. Each page should have tags which are created along with the page that describe the content on that page. I want to compare tags between pages to find the most relevant "related" pages on my site.

In case that isn't clear, what I am trying to explain would be very similar to the comma-separated "Tags" field just below this text box on the forum where "you may add 5 tag(s) to this thread."

What is the best way to implement this? Does anyone have a script that might get me started? Thanks in advance.

Monie · March 19th, 2008, 09:00 PM

Sorry, what are the Tags for? I mean what is the purpose?
Cheers...

MrQuestions · March 19th, 2008, 11:29 PM

The tags will be for showing "related pages," more specifically for determining whether or not the pages are actually related. At the bottom of each article displayed on my site, I would like to display a "related articles" section that shows links to 5 related articles determined by tags.

The article author will add tags to the article when it is created. on the .PHP page for an article, I would place code at the bottom that generates the list of links based on the tag algorithm that determines the 5 most relevant articles. I can think of a few different ways to do it, but I know there must be a lot of scripts out there that do already do this... so rather than trying to reinvent the wheel from scratch I thought I'd ask you guys for some suggestions. Thanks a lot for your reply!

dannyboy26 · March 20th, 2008, 09:06 PM

Lots of ways depending on exactly what you're after.

A simple method is to store all content for pages in a database. When the data gets added, run a script that counts which words/phrases occurs most often and store the top five in a separate field in the database, comma separated.

This then gets listed with each article, and turn each word into a key link. This then does a search and finds other articles in the database that have keywords in common.

That's just one of many possible ways of doing something along these lines...

Dan