Description

A class of objects designed to help aggregate calculations over an iterative computation. The aggregator consists of three objects. An environment to hold the values. A function that sets up an initial value the first time an object is seen. An aggregate function that increments the value of an object seen previously.

Details

This class is used to help aggregate different values over function calls. A very simple example is to use leave one out cross-validation for prediction. At each stage we first perform feature selection and then cross-validate. To keep track of how often each feature is selected we can use an aggregator. At the end of the cross-validation we can extract the names of the features chosen from aggenv.

Creating Objects

new('aggregator', aggenv = [environment], initfun = [function], aggfun = [function])

Slots

aggenv:

Object of class 'environment', holds the values between iterations

initfun:

Object of class 'function' specifies how to initialize the value for a name the first time it is encountered

aggfun:

Object of class 'function' used to increment (or perform any other function) on a name

Methods

aggenv(aggregator):

Used to access the environment of the aggregator

aggfun(aggregator):

Used to access the function that aggregates

initfun(aggregator):

Used to access the initializer function

See Also

Aggregate