‹‹ Back to SVS Home

17.1 Split-Prediction Methodology

17.1 Split-Prediction Methodology

There are a variety of methods that are used to either partition the data into subgroups or regress on a predictor variable. Furthermore, the statistics used to perform hypothesis testing depend on the distribution of the response. Two response distributions are supported–normal and binomial. Within each response distribution there are a number of splits or regressions that can be performed either at a tree node or on spreadsheet data.