module BC_Learning: sig end
Machine Learning stuff
Author(s): Hideo Bannai
val cross_validate : int ->
'a list ->
?sampler:(int -> 'a list -> ('a list * 'a list) list) ->
('a list -> 'b) -> ('b -> 'a list -> 'c) -> 'c list
cross_validate k data sampler learner evaluator
runs a k-cross validation on the data.
Each element of Data is of type 'a. The learner will
take a list of data, and output a hypothesis of type 'b.
evaluator will take the learned hypothesis, and data
and return some evaluation of type 'c.
The k evaluations are returned in a list.
sampler : A function to create cross validation data sets. The default is to use BC_Random.cross_sample_list.
val posneg_cross_validate : int ->
'a list ->
'a list ->
?sampler:(int -> 'a list -> ('a list * 'a list) list) ->
('a list -> 'a list -> 'b) -> ('b -> 'a list -> 'a list -> 'c) -> 'c list
posneg_cross_validate k posdata negdata sampler learner evaluator
similar to cross_validate:
runs a k-cross validation on the data, but the data consists of
two sets, which should be equally sampled.
Each element of Data is of type 'a. The learner will
take two lists of data, and output a hypothesis of type 'b.
evaluator will take the learned hypothesis, and two sets of data
and return some evaluation of type 'c.
The k evaluations are returned in a list.
sampler : A function to create cross validation data sets. The default is to use BC_Random.cross_sample_list.