Random-walk Metropolis-Hastings
Random-walk Metropolis-Hastings
(->entry choice-id value log-p cont)
Positional factory function for class anglican.rmh.entry.
Positional factory function for class anglican.rmh.entry.
(choice-id smp state)
returns a unique idenditifer for sample checkpoint and the updated state
returns a unique idenditifer for sample checkpoint and the updated state
(correct-log-weight state)
corrects log weight of a sample, setting it to 0 if the sample is in the support, -Infinity otherwise
corrects log weight of a sample, setting it to 0 if the sample is in the support, -Infinity otherwise
given the value and distribution of x_{nk}, returns an alternative proposal distribution (runtime/distribution protocol) which is centered around x_{nk} with some closeness factor sigma
given the value and distribution of x_{nk}, returns an alternative proposal distribution (runtime/distribution protocol) which is centered around x_{nk} with some closeness factor sigma
(get-kernel-proposal distribution value alpha sigma)
returns a kernel proposal distribution (runtime/distribution protocol) k( . | x_{nk}) given the the value and the distribution of x_{nk}. The kernel is constructed by choosing an alternative proposal distribution using a biased coin flip with the success probability alpha.
returns a kernel proposal distribution (runtime/distribution protocol) k( . | x_{nk}) given the the value and the distribution of x_{nk}. The kernel is constructed by choosing an alternative proposal distribution using a biased coin flip with the success probability alpha.
(get-log-kappa state)
computes log probability arising from the transition kappa: k(x'{nk} | x{nk})
computes log probability arising from the transition kappa: k(x'_{nk} | x_{nk})
(get-log-new-val-from-prior state)
computes log probability of the new choice x'_{nk} pretending it was sampled under prior
computes log probability of the new choice x'_{nk} pretending it was sampled under prior
(get-log-retained-probability state)
computes log probability of retained random choices
computes log probability of retained random choices
(map->entry m__7585__auto__)
Factory function for class anglican.rmh.entry, taking a map of keywords to field values.
Factory function for class anglican.rmh.entry, taking a map of keywords to field values.
(next-state state entry)
(next-state state entry update)
produces next state given current state and the trace entry to resample
produces next state given current state and the trace entry to resample
(prev-state state next-state entry)
(prev-state state next-state entry update)
produces previous state given the current and the next state and the resampled entry by re-attaching new rdb to the original state
produces previous state given the current and the next state and the resampled entry by re-attaching new rdb to the original state
(rdb trace)
creates random database from trace
creates random database from trace
(record-choice state choice-id value log-p cont)
records random choice in the state
records random choice in the state
(utility state)
computes state utility, used to determine the acceptance log-probability as (next-utility - prev-utility)
computes state utility, used to determine the acceptance log-probability as (next-utility - prev-utility)
cljdoc is a website building & hosting documentation for Clojure/Script libraries
× close