^inductive programming^

Conversion Mappings

Useful conversion functions map from one class of statistical model to another.

They directly implement basic probability rules in the programming language.

They also increase the generality of models and estimators.

users.monash.edu/~lloyd/Seminars/2005-II/Conversion/index.shtml  
Conversion function Parameter(s) Result
model2timeSeries a model of data-space 'ds' the trivial time-series model of ds
model2functionModel a model of data-space 'ds' the trivial function-model from an input-space, 'is', to output-space ds
condition
(Bayes's thm)
a function-model from 'is' to 'os', and a model of 'is' joint model of (is,os)
functionModel2model a function-model from 'is' to 'os' a model of (is,os)
where 'is' is "common knowledge"
functionModel2timeSeries a function-model from [ds] to ds a time-series model of ds
timeSeries2model a time-series model of ds
(& a model of lengths (Int))
a model of [ds]
timeSeries2functionModel a time-series model of ds a function-model from [ds] to ds

See: L. Allison. Models for Machine Learning and Data Mining in Functional Programming, J. Functional Programming (JFP), 15(1), pp.15-32, January 2005, and also [II(click)].

PS. On Haskell

A value has a type (::),  a type can be an instance of one or more (type-)classes.
e.g. True :: Bool,  Bool is an instance of Read and Show.
 
[t]   -- list type, where 't' is the element type.
i -> v   -- function type, input type 'u', result type 'v'.