I am trying to code an algorithm that acts a bit like Amazon\'s \"People who bought this also bought\".
The difference between the two is that mine is just counting the
I would make one improvement to your idea. When you're figuring out the cliques that go together and deciding which make the strongest relationships you should add a weight against each connection. The easiest way to calculate the weight would be to see how many people who looked at product X also looked at Y. The more views, the stronger the relationship.