Everybody who writes about Data Mining has to get around to formally defining it at one point or another.
Data Mining is the search of data, accumulated during the normal course of doing business, in order to find and confirm the existence of previously unknown relationships that can produce positive and verifiable outcomes through the deployment of predictive models when applied to new data.
Below are the key features. One or more are often missing from definitions I come across in my reading:
– The mined data is not new
– The data that can answer the question was not collected solely to mine
– The miner is not testing hypotheses or known relationships
– The relationships must be verifiable
– The resulting models must be useful
– The resulting models must work on new data