I'm trying to find a way to classify stock price data that has been sampled at random, uncorrelated, time periods. While looking for an algorithm to help me use a correlation matrix for classification, I came across stock correlation networks. Unfortunately, I don't know anything about these networks, so I'm wondering if there's a specific network method someone could point me to that would be particularly good for my purpose.
Prior to coming across correlation networks my basic idea was to set a correlation threshold and find or make an algorithm to create sets which use a secondary criterion for stocks that ended up in multiple or no sets. It sounds like correlation networks use a very similar, more complex idea, but if there's a better, simpler, way I'm definitely open to suggestions.