Data Mining in Networks

Data mining algorithms learn knowledge of the type necessary to make these inferences. Data mining algorithms learn such knowledge based on data for which the correct prediction is known (e.g., databases containing many past cases of money laundering and many cases of legitimate business activity).

Given such data, data mining algorithms discover knowledge about how to make the correct inference. Some data mining algorithms construct the type of logical rules shown in the previous slides, and others construct models represented as equations, classification trees, or graphical networks of associations.

This knowledge is then used to make inferences with data for which the correct prediction is not known. For example, the rules about money laundering might be applied to data drawn from ongoing cases from law enforcement agencies and data on large currency transactions gathered by Treasury.