Estimating Functions of Mixed Ordinal and Categorical Variables Using Adaptive Splines

摘要: Abstract : Multivariate adaptive regression splines (MARS) is a methodology for nonparametrically estimating (and interpreting) general functions of high-dimensional argument given (usually noisy) data. Its basic underlying assumption that the function to be estimated locally relatively smooth where smoothness adaptively defined depending on local characteristics function. The usual definitions do not apply variables assume unorderable categorical values. After brief review MARS strategy ordinal variables, alternative concepts appropriate are introduced. These lead procedures can estimate and interpret many as well those involving (many) mixed variables. They also provide natural mechanism modeling predicting in presence missing predictor values (ordinal or categorical).