mml

Class Adaptive.M

java.lang.Object

la.la.Value

mml.Model

mml.UPModel.M

mml.Discretes.M

mml.Discretes.Bounded.M

mml.Adaptive.M

All Implemented Interfaces:

java.lang.Comparable<Value>

Enclosing class:

Adaptive

public class Adaptive.M
extends Discretes.Bounded.M

Model Mdl should be enough for most purposes but here is its class, fully (trivially) parameterised (M). Also see the UnParameterised Model. Note the remarks on nlLH(ss) and nlPr_n(n).

Nested Class Summary

Nested classes/interfaces inherited from class mml.Model

Model.Defaults, Model.Transform

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 inherited from class mml.Model

msg1, msg2, sp

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
M(double msg1, double msg2, Value sp) Requires msg1=0, sp=triv; msg2 may be > 0 for fair competition on training data.

Method Summary

Modifier and Type	Method and Description
double	nlLH(Value ss) The negative log likelihood of data-set 'ds', where statistics ss = stats(ds).
double	nlPr_n(int n) The negative log probability of datum int 'n', but pay careful attention to the remarks on nlLH(ss).
double	nlPr(Value d) The negative log probability of Discrete datum 'd', but pay careful attention to the remarks on nlLH(ss).
int	random_n() Not implemented, arguably it is not possible (do not think that using random(1) truly gets around this).
Vector	random(int n) Uses randomSeries() to sample 'n' random homogeneous Values.
Series.Discrete	randomSeries() Return a Series.Discrete which is random and homogeneous: Sample probabilities 'prs[]' from Dirichlet.M(alpha) and then repeatedly sample Ints from MultiState.M(prs), i.e., the Series is random and homogeneous.

Methods inherited from class mml.Discretes.Bounded.M

asUPModel, bounds, lwb_n, lwb, random, type, upb_n, upb

Methods inherited from class mml.Discretes.M

pr_n, pr, shifted

Methods inherited from class mml.UPModel.M

asGiven, asGiven, stats, stats, toString

Methods inherited from class mml.Model

asEstimator, m1m2sp, msg, msg1, msg1bits, msg2, msg2bits, msgBits, nl2LH, nl2Pr, statParams, stats, stats, sumNlPr, transform, zeroTriv

Methods inherited from class la.la.Value

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

Methods inherited from class java.lang.Object

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

Constructor Detail

M

public M(double msg1,
         double msg2,
         Value sp)

Requires msg1=0, sp=triv; msg2 may be > 0 for fair competition on training data.

Method Detail

nlPr

public double nlPr(Value d)

The negative log probability of Discrete datum 'd', but pay careful attention to the remarks on nlLH(ss). ?Arguably nlPr(d) should be an Unsupported Operation?

Overrides:

nlPr in class Discretes.M

nlPr_n

public double nlPr_n(int n)

The negative log probability of datum int 'n', but pay careful attention to the remarks on nlLH(ss). ?Arguably nlPr_n(n) should be an Unsupported Operation?

Specified by:

nlPr_n in class Discretes.M

nlLH

public double nlLH(Value ss)

The negative log likelihood of data-set 'ds', where statistics ss = stats(ds). The running frequency "counter" for Value v_k is, in effect, initialised to α_k. Note that nlLH(ss) may be quite different to ∑_i nlPr(ds_i) because the latter does not, and cannot, "adapt".

Specified by:

nlLH in class Model

random_n

public int random_n()

Not implemented, arguably it is not possible (do not think that using random(1) truly gets around this). For a data-set, random(n) and randomSeries() are fine.

Overrides:

random_n in class Discretes.Bounded.M

random

public Vector random(int n)

Uses randomSeries() to sample 'n' random homogeneous Values.

Overrides:

random in class Discretes.Bounded.M

randomSeries

public Series.Discrete randomSeries()

Return a Series.Discrete which is random and homogeneous: Sample probabilities 'prs[]' from Dirichlet.M(alpha) and then repeatedly sample Ints from MultiState.M(prs), i.e., the Series is random and homogeneous. Don't forget to advance()!

Overrides:

randomSeries in class Discretes.Bounded.M