I can’t resist a good pattern. Patchwork quilts, the golden ratio, diagramming sentences and classification systems… love them all. I find it reassuring that there can be order in perceived states of chaos — this is how I justify the way my office looks. Oh sure, it appears that things are out of order, but I promise, I can find the paper you’re asking me for.
Science magazine describes in their latest issue that scientists have developed a new program, called MINE, that can find patterns in data scientists don’t know how to look for. They describe it as:
“useful for identifying and characterizing structure in data.”
I don’t pretend to understand even a fraction of the math involved (pun absolutely intended)… but my sense is that the program applies different ways of analyzing patterns (sequence, exponential, one-step-back-two-steps-forward, etc) and see what fits. Like the Cinderella method — if both glass slippers fit, then we have a winner!
Maybe there are sociological applications as well beyond financial modeling and genome sequencing. Like, let’s put the data points of my dating life into the program and see what patterns emerge, because while I’m fairly certain there’s a pattern, I just haven’t been able to figure it out.
Then again, can’t scientists agree that there are some data sets that are truly random?