Generic N-fold data partitioner.
Given a dataset with N chunks, with cvtype = 1 (which is default), it would generate N partition sets, where each chunk is sequentially taken out (with replacement) to form a second partition, while all other samples together form the first partition. Example, if there are 4 chunks, partition sets for cvtype = 1 are:
[[1, 2, 3], [0]]
[[0, 2, 3], [1]]
[[0, 1, 3], [2]]
[[0, 1, 2], [3]]
If cvtype>1, then all possible combinations of cvtype number of chunks are taken out, so for cvtype = 2 in previous example yields:
[[2, 3], [0, 1]]
[[1, 3], [0, 2]]
[[1, 2], [0, 3]]
[[0, 3], [1, 2]]
[[0, 2], [1, 3]]
[[0, 1], [2, 3]]
Note that the “taken-out” partition is always labeled ‘2’ while the remaining elements are labeled ‘1’.
If cvtype is a float in the range from 0 to 1, it specifies the ratio of present unique values to be taken.
If cvtype is large enough generating prohibitively large number of combinations, provide count to limit number of combinations and provide selection_strategy = ‘random’.
Notes
Available conditional attributes:
(Conditional attributes enabled by default suffixed with +)
Parameters : | cvtype : int, float
enable_ca : None or list of str
disable_ca : None or list of str
count : None or int
selection_strategy : str
attr : str
space : str
postproc : Node instance, optional
descr : str
|
---|