Data mining refers to the process of actively looking at large sets of data for any patterns (correlations). But since random data are clumpy, we should expect to see accidental correlations even when there is no real underlying phenomenon causing the correlation.
