Package org.ptolemy.machineLearning.hsmm

Class HSMMTimeAwareMultinomialEstimator

    • Field Detail

      • timestamps

        public TypedIOPort timestamps
        Array of observation timestamps as UNIX timestamps.

      • transitionMatrixEstimationMethod

        public Parameter transitionMatrixEstimationMethod
        Transition Matrix partitioning options. Force self transition asserts a self transition with probability 1 if no information has been learned for the state. Interpolate: Assigns uniform probabilities to any state that has Hamming distance less than two to the current state.

      • empiricalStartTimes

        public TypedIOPort empiricalStartTimes
        Array of estimated probability transition matrices for each hour.

      • NUM_CATEGORIES

        protected static final int NUM_CATEGORIES
        Number of partitions in the probability transition matrix.
        See Also:
        Constant Field Values

      • INTERPOLATE

        protected static final java.lang.String INTERPOLATE
        Completion strategy for A set to interpolation, that is, a uniform distribuition on all states that have Hamming distance <= 1 to the binary representation of the source state.
        See Also:
        Constant Field Values

      • FORCE_SELF

        protected static final java.lang.String FORCE_SELF
        Completion strategy for A set to forcing a self-transition.
        See Also:
        Constant Field Values

      • FORCE_ZERO

        protected static final java.lang.String FORCE_ZERO
        Completion strategy for A set to forcing a transition to state 0.
        See Also:
        Constant Field Values

      • NO_ACTION

        protected static final java.lang.String NO_ACTION
        No completion strategy.
        See Also:
        Constant Field Values

      • SELF_AND_ZERO

        protected static final java.lang.String SELF_AND_ZERO
        Completion strategy for A set to self and zero transitions with equal probability.
        See Also:
        Constant Field Values

      • At

        protected double[][][] At
        Time-dependent transition probability matrix.

      • Atlearned

        protected double[][][] Atlearned
        The learned transition probability matrix: before completion strategy is applied.

      • incompleteCategories

        protected java.util.Set<int[]> incompleteCategories
        Hour categories for which At has not enough information.

      • _hourOfDay

        protected int[] _hourOfDay
        hour of day for input observations.

    • Method Detail

      • attributeChanged

        public void attributeChanged​(Attribute attribute)
                      throws IllegalActionException
        Description copied from class: NamedObj
        React to a change in an attribute. This method is called by a contained attribute when its value changes. In this base class, the method does nothing. In derived classes, this method may throw an exception, indicating that the new attribute value is invalid. It is up to the caller to restore the attribute to a valid value if an exception is thrown.
        Overrides:
        attributeChanged in class HSMMMultinomialEstimator
        Parameters:
        attribute - The attribute that changed.
        Throws:
        IllegalActionException - If the change is not acceptable to this container (not thrown in this base class).

      • _learnAt

        protected void _learnAt()
        Learn the transition probability matrix for each hour, from timestamped data.

      • _calculateTransitionScheme

        protected void _calculateTransitionScheme​(java.lang.String method,
                                          int category)
        Compute missing transition probabilities in the hourly transition matrix estimates.
        Parameters:
        method - Completion strategy
        category - Source category