Question

Why data mining would be a problem for predictive models?

Answer 1

Predictive models are based on data of the past provided by data mining. Data mining can be a problem for predictive model because data mining forms the base of the model. If the data provided for data mining are not sufficient, manipulated, proliferated , then the data mining is not done properly and provides vague patterns and relationships, which when used in predictive model will distort the model.

Data mining requires mining of huge amount of data. So it is possible that different approaches may be used , which may provide different results. This in turn will again make predictive model confusing.