mirror of
https://github.com/hanzoai/jube-fork.git
synced 2026-10-30 18:33:33 +00:00
richard-churchman
cc0a2bee21
Included open-source Accord.net library code natively to the solution given the deprecated state of that project. Accord.net will be supported inside Jube for own needs (such as upgrading to .Net 7+). Accord.net libraries modified to compile to .Net 6. Modified Accord.net libraries to better support Newtonsoft Json.net serialisation. Included annotations in the Accord.net libraries to house important dependency statistics. Serialised json can be used independently of Jube to recall a Neural Network. Removal of Binary Serializer totally as a precursor to the projects support of .Net 7+. Extensive bug fixing in crucial functions of Neural Network training and recall. While this part of the system remains in Alpha, it is substantially more fit for purpose.
234 lines
8.3 KiB
C#
234 lines
8.3 KiB
C#
// Accord Statistics Library
|
|
// The Accord.NET Framework
|
|
// http://accord-framework.net
|
|
//
|
|
// Copyright © César Souza, 2009-2017
|
|
// cesarsouza at gmail.com
|
|
//
|
|
// This library is free software; you can redistribute it and/or
|
|
// modify it under the terms of the GNU Lesser General Public
|
|
// License as published by the Free Software Foundation; either
|
|
// version 2.1 of the License, or (at your option) any later version.
|
|
//
|
|
// This library is distributed in the hope that it will be useful,
|
|
// but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
|
|
// Lesser General Public License for more details.
|
|
//
|
|
// You should have received a copy of the GNU Lesser General Public
|
|
// License along with this library; if not, write to the Free Software
|
|
// Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
|
|
//
|
|
|
|
namespace Accord.Statistics.Models.Markov.Topology
|
|
{
|
|
using System;
|
|
using Accord.Math;
|
|
using System.Data;
|
|
using Accord.Compat;
|
|
|
|
/// <summary>
|
|
/// Custom Topology for Hidden Markov Model.
|
|
/// </summary>
|
|
///
|
|
/// <remarks>
|
|
/// <para>
|
|
/// An Hidden Markov Model Topology specifies how many states and which
|
|
/// initial probabilities a Markov model should have. Two common topologies
|
|
/// can be discussed in terms of transition state probabilities and are
|
|
/// available to construction through the <see cref="Ergodic"/> and
|
|
/// <see cref="Forward"/> classes implementing the <see cref="ITopology"/>
|
|
/// interface.</para>
|
|
///
|
|
/// <para>Topology specification is important with regard to both learning and
|
|
/// performance: A model with too many states (and thus too many settable
|
|
/// parameters) will require too much training data while an model with an
|
|
/// insufficient number of states will prohibit the HMM from capturing subtle
|
|
/// statistical patterns.</para>
|
|
///
|
|
/// <para>This custom implementation allows for arbitrarily specification of
|
|
/// the state transition matrix and initial state probabilities for
|
|
/// <see cref="IHiddenMarkovModel">hidden Markov models</see>.</para>
|
|
///
|
|
/// </remarks>
|
|
///
|
|
/// <seealso cref="HiddenMarkovModel"/>
|
|
/// <seealso cref="Accord.Statistics.Models.Markov.Topology.ITopology"/>
|
|
/// <seealso cref="Accord.Statistics.Models.Markov.Topology.Ergodic"/>
|
|
/// <seealso cref="Accord.Statistics.Models.Markov.Topology.Forward"/>
|
|
///
|
|
[Serializable]
|
|
public class Custom : ITopology
|
|
{
|
|
|
|
private int states;
|
|
private double[] pi;
|
|
private double[,] A;
|
|
|
|
/// <summary>
|
|
/// Creates a new custom topology with user-defined
|
|
/// transition matrix and initial state probabilities.
|
|
/// </summary>
|
|
///
|
|
/// <param name="initial">The initial probabilities for the model.</param>
|
|
/// <param name="transitions">The transition probabilities for the model.</param>
|
|
///
|
|
public Custom(double[][] transitions, double[] initial)
|
|
: this(transitions, initial, false) { }
|
|
|
|
/// <summary>
|
|
/// Creates a new custom topology with user-defined
|
|
/// transition matrix and initial state probabilities.
|
|
/// </summary>
|
|
///
|
|
/// <param name="initial">The initial probabilities for the model.</param>
|
|
/// <param name="transitions">The transition probabilities for the model.</param>
|
|
/// <param name="logarithm">Set to true if the passed transitions are given
|
|
/// in log-probabilities. Default is false (given values are probabilities).</param>
|
|
///
|
|
public Custom(double[][] transitions, double[] initial, bool logarithm)
|
|
: this(transitions.ToMatrix(), initial, logarithm) { }
|
|
|
|
/// <summary>
|
|
/// Creates a new custom topology with user-defined
|
|
/// transition matrix and initial state probabilities.
|
|
/// </summary>
|
|
///
|
|
/// <param name="initial">The initial probabilities for the model.</param>
|
|
/// <param name="transitions">The transition probabilities for the model.</param>
|
|
///
|
|
public Custom(double[,] transitions, double[] initial)
|
|
: this(transitions, initial, false) { }
|
|
|
|
/// <summary>
|
|
/// Creates a new custom topology with user-defined
|
|
/// transition matrix and initial state probabilities.
|
|
/// </summary>
|
|
///
|
|
/// <param name="initial">The initial probabilities for the model.</param>
|
|
/// <param name="transitions">The transition probabilities for the model.</param>
|
|
/// <param name="logarithm">Set to true if the passed transitions are given
|
|
/// in log-probabilities. Default is false (given values are probabilities).</param>
|
|
///
|
|
public Custom(double[,] transitions, double[] initial, bool logarithm)
|
|
{
|
|
if (transitions == null)
|
|
{
|
|
throw new ArgumentNullException("transitions");
|
|
}
|
|
|
|
if (initial == null)
|
|
{
|
|
throw new ArgumentNullException("initial");
|
|
}
|
|
|
|
if (transitions.GetLength(0) != transitions.GetLength(1))
|
|
{
|
|
if (transitions.GetLength(0) != transitions.GetLength(1))
|
|
throw new ArgumentException(
|
|
"Transition matrix should be square.",
|
|
"transitions");
|
|
}
|
|
|
|
if (initial.Length != transitions.GetLength(0))
|
|
{
|
|
throw new ArgumentException(
|
|
"Initial probabilities should have the same length as the number of states in the transition matrix.",
|
|
"initial");
|
|
}
|
|
|
|
this.states = transitions.GetLength(0);
|
|
|
|
for (int i = 0; i < states; i++)
|
|
for (int j = 0; j < states; j++)
|
|
if (Double.IsNaN(transitions[i, j]))
|
|
throw new ArgumentOutOfRangeException("transitions");
|
|
|
|
if (logarithm)
|
|
{
|
|
transitions = transitions.Exp();
|
|
initial = initial.Exp();
|
|
}
|
|
|
|
this.A = transitions;
|
|
this.pi = initial;
|
|
|
|
double initialSum = 0;
|
|
for (int i = 0; i < pi.Length; i++)
|
|
initialSum += pi[i];
|
|
|
|
if (!initialSum.IsEqual(1, 1e-10))
|
|
{
|
|
throw new ArgumentException("Initial probabilities do not sum up to one.",
|
|
"initial");
|
|
}
|
|
|
|
for (int i = 0; i < states; i++)
|
|
{
|
|
double stateSum = 0;
|
|
for (int j = 0; j < states; j++)
|
|
stateSum += transitions[i, j];
|
|
|
|
if (!stateSum.IsEqual(1, 1e-10))
|
|
{
|
|
throw new ArgumentException("State probabilities for state " + i +
|
|
" do not sum up to one.", "transitions");
|
|
}
|
|
}
|
|
}
|
|
|
|
|
|
/// <summary>
|
|
/// Gets the number of states in this topology.
|
|
/// </summary>
|
|
///
|
|
public int States
|
|
{
|
|
get { return states; }
|
|
set { throw new ReadOnlyException("Property is read-only."); }
|
|
}
|
|
|
|
/// <summary>
|
|
/// Gets the initial state probabilities.
|
|
/// </summary>
|
|
///
|
|
public double[] Initial
|
|
{
|
|
get { return pi; }
|
|
}
|
|
|
|
/// <summary>
|
|
/// Gets the state-transitions matrix.
|
|
/// </summary>
|
|
///
|
|
public double[,] Transitions
|
|
{
|
|
get { return A; }
|
|
}
|
|
|
|
|
|
|
|
/// <summary>
|
|
/// Creates the state transitions matrix and the
|
|
/// initial state probabilities for this topology.
|
|
/// </summary>
|
|
///
|
|
public int Create(bool logarithm, out double[,] transitionMatrix, out double[] initialState)
|
|
{
|
|
if (logarithm)
|
|
{
|
|
transitionMatrix = Elementwise.Log(A);
|
|
initialState = Elementwise.Log(pi);
|
|
}
|
|
else
|
|
{
|
|
transitionMatrix = (double[,])A.Clone();
|
|
initialState = (double[])pi.Clone();
|
|
}
|
|
|
|
return states;
|
|
}
|
|
|
|
}
|
|
}
|