Predictive modeling is a statistical technique that makes use of patterns and relationships in data to predict future outcomes. There are a number of predictive modeling methods. Each model consists of a number of variables/predictors that are likely to predict future results. Typically the objective is to determine which of these variables are the best predictors.
For example, we may want to try and better understand what demographics from an existing database of students who go to a college are the best predictors of their likelihood of dropping out before graduation. Once we know this, incoming freshman who exhibit these predictors can be identified and programs put in place to increase the chances of retaining them.
Predictive modeling can be complex. However, this complexity is simplified by making use of statistical software, such as SPSS Modeler, to build and run predictive models. Below are a few selected videos that offer a good overview of predictive modeling using SPSS Modeler.
This last one is a webinar, so it is a little on the long side. The core of the presentation that deals with the more applicable aspects of predictive modeling starts at about minute 34.