mml

Class Estimator

java.lang.Object

la.la.Value

la.la.Function

la.la.Function.Native

mml.Estimator

All Implemented Interfaces:

java.lang.Comparable<Value>

Direct Known Subclasses:

UPModel.Est

public abstract class Estimator
extends Function.Native

An Estimator estimates (fits) a fully parameterised Model to a data-set, thus solving the inference problem posed by some UnParameterised Model. Most Estimators will be instances of UPModel.Est, UPFunctionModel.Est or UPSeriesModel.Est. Note that an Estimator can do all that a UPModel can do, sp2Model(m1,m2,sp) & stats(ds), and in addition ds2Model(ds) & ss2Model(ss). In general these last two depend on the Estimator's own parameter(s) which may, for example, control a prior.

Nested Class Summary

Nested classes/interfaces inherited from class la.la.Function.Native

Function.Native.WithInverse

Nested classes/interfaces inherited from class la.la.Function

Function.Cts2Cts, Function.Cts2Cts2Cts, Function.CtsD2CtsD, Function.HasInverse, Function.Native, Function.Native2, Function.Native3

Nested classes/interfaces inherited from class la.la.Value

Value.Atomic, Value.Bool, Value.Char, Value.Chars, Value.Cts, Value.Defer, Value.Discrete, Value.Enum, Value.Inc_Or, Value.Int, Value.Lambda, Value.List, Value.Maybe, Value.Option, Value.Real, Value.Scannable, Value.Structured, Value.Triv, Value.Tuple

Field Summary

Fields

Modifier and Type	Field and Description
Value	params The Estimator's parameters, notably the parameters of a prior on the estimated-Model's statistical parameters.

Fields inherited from class la.la.Value

C, comparator, CR, E, eight, ffalse, five, fiveR, four, fourR, half, negCR, negOne, negOneR, negTenR, nilVal, nine, None, one, oneR, PI, PIby2, point01, point1, point5, point9, quarter, seven, six, ten, tenR, three, threeR, triv, ttrue, two, twoR, zero, zeroR

Constructor Summary

Constructors

Constructor and Description
Estimator(Value params) An Estimator may have parameters, notably parameters of a prior.

Method Summary

Modifier and Type	Method and Description
Model	apply(Value ds) Apply, 'this' Estimator to a data-set, ds, that is, return the Model ds2Model(ds).
UPModel	asUPModel() In some context, it might(?) be necessary to treat 'this' Estimator as an UnParameterised Model, upm, where upm.estimator(triv) always returns 'this' Estimator.
abstract Value	defnParams() Return problem-defining parameter, for example, by UPModel.defnParams().
Model	ds2Model(Vector ds) Given a data-set, ds, estimate a Model by mdl = ss2Model(stats(ds)).
Value.Tuple	ds2ModelSp(Vector ds) Given a data-set, ds, return ss2ModelSp(stats(ds)); read ss2ModelSp(ss) carefully!
abstract Model	sp2Model(double m1, double m2, Value sp) Given part 1 and part 2 message lengths, m1 and m2, and statistical parameter(s), sp, return a fully parameterised Model.
abstract Model	ss2Model(Value ss) Given sufficient statistics, ss=stats(ds), of a data-set, ds, estimate a fully parameterised Model, ss2Model(ss).
Value.Tuple	ss2ModelSp(Value ss) Given stats, ss, estimate a Model and parameters; this default is (mdl, sp), where mdl=ss2Model(ss) and sp=mdl.statParams() (usually).
abstract Value	stats(boolean add, Value ss0, Value ss1) Combine statisticses ss0 and ss1, '+' if add=true otherwise '-'.
Value	stats(boolean add, Value ss0, Vector ds, int lo, int hi) Add (remove) statistics for ds.[lo,hi) to (from) statistics ss0.
Value	stats(Vector ds) stats(0,ds.nElts()).
abstract Value	stats(Vector ds, int lo, int hi) Given a data-set 'ds', return sufficient statistics for elements [lo, hi), lo inclusive to hi exclusive, ss = stats(ds,lo,hi).
java.lang.String	toString() Return a String representation of 'this' Estimator.

Methods inherited from class la.la.Function

bOp, fromBop, fromUop, main, type, uOp

Methods inherited from class la.la.Value

AoM, bOp, compareTo, cons, cts, elt, errMsg, error, force, isTuple, just, n, nElts, nlAoM, pair, print, quad, real, RTE, triple, tuple, UOE, x

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, wait, wait, wait

Field Detail

params

public final Value params

The Estimator's parameters, notably the parameters of a prior on the estimated-Model's statistical parameters. Note that 'params' may be trivial.

Constructor Detail

Estimator

public Estimator(Value params)

An Estimator may have parameters, notably parameters of a prior.

Method Detail

apply

public final Model apply(Value ds)

Apply, 'this' Estimator to a data-set, ds, that is, return the Model ds2Model(ds). Also see Function.apply(la.la.Value).

Specified by:

apply in class Function.Native

defnParams

public abstract Value defnParams()

Return problem-defining parameter, for example, by UPModel.defnParams().

sp2Model

public abstract Model sp2Model(double m1,
                               double m2,
                               Value sp)

Given part 1 and part 2 message lengths, m1 and m2, and statistical parameter(s), sp, return a fully parameterised Model. Also see UPModel.sp2Model(m1,m2,sp).

stats

public Value stats(Vector ds)

stats(0,ds.nElts()). There is unlikely to be any need for a subclass to override stats(ds).

stats

public abstract Value stats(Vector ds,
                            int lo,
                            int hi)

Given a data-set 'ds', return sufficient statistics for elements [lo, hi), lo inclusive to hi exclusive, ss = stats(ds,lo,hi). 'ss' is used to estimate a Model as in ss2Model(ss), say. More on stats [here].

stats

public abstract Value stats(boolean add,
                            Value ss0,
                            Value ss1)

Combine statisticses ss0 and ss1, '+' if add=true otherwise '-'.

stats

public Value stats(boolean add,
                   Value ss0,
                   Vector ds,
                   int lo,
                   int hi)

Add (remove) statistics for ds.[lo,hi) to (from) statistics ss0. There is unlikely to be any need for a subclass to override stats(add,ss0,ds,lo,hi).

ds2ModelSp

public final Value.Tuple ds2ModelSp(Vector ds)

Given a data-set, ds, return ss2ModelSp(stats(ds)); read ss2ModelSp(ss) carefully!

ds2Model

public final Model ds2Model(Vector ds)

Given a data-set, ds, estimate a Model by mdl = ss2Model(stats(ds)). Also see ds2ModelSp(ds)!

ss2ModelSp

public Value.Tuple ss2ModelSp(Value ss)

Given stats, ss, estimate a Model and parameters; this default is (mdl, sp), where mdl=ss2Model(ss) and sp=mdl.statParams() (usually). However, there is at least one exception – Model.asUPModel()'s estimator(()) – which returns (mdl, triv). The distinction ss2ModelSp:ss2Model, and ds2ModelSp:ds2Model, can matter to a compound UnParameterised Model handed a Model.asUPModel().

ss2Model

public abstract Model ss2Model(Value ss)

Given sufficient statistics, ss=stats(ds), of a data-set, ds, estimate a fully parameterised Model, ss2Model(ss). Also see ss2Model(ss)!

toString

public java.lang.String toString()

Return a String representation of 'this' Estimator.

Overrides:

toString in class Function

asUPModel

public UPModel asUPModel()

In some context, it might(?) be necessary to treat 'this' Estimator as an UnParameterised Model, upm, where upm.estimator(triv) always returns 'this' Estimator. Also see Model.asUPModel().