Rev Language Reference
mvRandomCategoryWalk - Random walk on vector of naturals (e.g., from multinomial distribution) keeping
the sum of the vector constant.
This random walk proposal picks a random index of a vector. Then, it picks
a random neighbor, either one index left or one index right. Then it decreases
the current value at the chosen index by one, and increases the value of the
neighbor by one.
Such a move is useful for values drawn from a multinomial distribution because
it keeps the total constant.
Usage
Arguments
| x : | Natural[] (<stochastic> pass by reference) |
| The variable on which this move operates. | |
| weight : | RealPos (pass by value) |
| The weight determines the relative frequency with which this move will be attempted. For details, see the description of the 'moveschedule' parameter on the documentation page for 'mcmc()'. | |
| Default : 1 |
Example
# create a vector of moves
moves = VectorMoves()
# draw from a multinomial distribution
num_e_prior <- simplex(rep(1, 6))
number_events_pi ~ dnMultinomial(p=num_e_prior, size=20)
# place a move on the draw
moves.append( mvRandomCategoryWalk(x=number_events_pi, weight=5) )
Edit this page