U Kv f.@svdZddlmZddlZddlmZGdddZGdddZGd d d Z Gd d d Z d dZ GdddZ dS)a! Utilities for cross validation. taken from scikits.learn # Author: Alexandre Gramfort , # Gael Varoquaux # License: BSD Style. # $Id$ changes to code by josef-pktd: - docstring formatting: underlines of headers )lrangeN) combinationsc@s(eZdZdZddZddZddZdS) LeaveOneOutzs Leave-One-Out cross validation iterator: Provides train/test indexes to split data in train test sets cCs ||_dS)a9 Leave-One-Out cross validation iterator: Provides train/test indexes to split data in train test sets Parameters ---------- n: int Total number of elements Examples -------- >>> from scikits.learn import cross_val >>> X = [[1, 2], [3, 4]] >>> y = [1, 2] >>> loo = cross_val.LeaveOneOut(2) >>> for train_index, test_index in loo: ... print "TRAIN:", train_index, "TEST:", test_index ... X_train, X_test, y_train, y_test = cross_val.split(train_index, test_index, X, y) ... print X_train, X_test, y_train, y_test TRAIN: [False True] TEST: [ True False] [[3 4]] [[1 2]] [2] [1] TRAIN: [ True False] TEST: [False True] [[1 2]] [[3 4]] [1] [2] N)n)selfrrG/tmp/pip-unpacked-wheel-2v6byqio/statsmodels/sandbox/tools/cross_val.py__init__szLeaveOneOut.__init__ccsB|j}t|D].}tj|td}d||<t|}||fVqdSNZdtypeT)rrangenpzerosbool logical_not)rri test_index train_indexrrr__iter__8s   zLeaveOneOut.__iter__cCsd|jj|jj|jfSNz %s.%s(n=%i) __class__ __module____name__rrrrr__repr__AszLeaveOneOut.__repr__Nrr __qualname____doc__r rrrrrrrs rc@s(eZdZdZddZddZddZdS) LeavePOutzq Leave-P-Out cross validation iterator: Provides train/test indexes to split data in train test sets cCs||_||_dS)aV Leave-P-Out cross validation iterator: Provides train/test indexes to split data in train test sets Parameters ---------- n: int Total number of elements p: int Size test sets Examples -------- >>> from scikits.learn import cross_val >>> X = [[1, 2], [3, 4], [5, 6], [7, 8]] >>> y = [1, 2, 3, 4] >>> lpo = cross_val.LeavePOut(4, 2) >>> for train_index, test_index in lpo: ... print "TRAIN:", train_index, "TEST:", test_index ... X_train, X_test, y_train, y_test = cross_val.split(train_index, test_index, X, y) TRAIN: [False False True True] TEST: [ True True False False] TRAIN: [False True False True] TEST: [ True False True False] TRAIN: [False True True False] TEST: [ True False False True] TRAIN: [ True False False True] TEST: [False True True False] TRAIN: [ True False True False] TEST: [False True False True] TRAIN: [ True True False False] TEST: [False False True True] N)rp)rrr rrrr PszLeavePOut.__init__ccsX|j}|j}tt||}|D]4}tj|td}d|t|<t|}||fVqdSr ) rr rrr rrarrayr)rrr combidxrrrrrrps zLeavePOut.__iter__cCsd|jj|jj|j|jfS)Nz%s.%s(n=%i, p=%i))rrrrr rrrrr{s zLeavePOut.__repr__NrrrrrrJs  rc@s(eZdZdZddZddZddZdS) KFoldzm K-Folds cross validation iterator: Provides train/test indexes to split data in train test sets cCs@|dksttd||ks0ttd||f||_||_dS)a K-Folds cross validation iterator: Provides train/test indexes to split data in train test sets Parameters ---------- n: int Total number of elements k: int number of folds Examples -------- >>> from scikits.learn import cross_val >>> X = [[1, 2], [3, 4], [1, 2], [3, 4]] >>> y = [1, 2, 3, 4] >>> kf = cross_val.KFold(4, k=2) >>> for train_index, test_index in kf: ... print "TRAIN:", train_index, "TEST:", test_index ... X_train, X_test, y_train, y_test = cross_val.split(train_index, test_index, X, y) TRAIN: [False False True True] TEST: [ True True False False] TRAIN: [ True True False False] TEST: [False False True True] Notes ----- All the folds have size trunc(n/k), the last one has the complementary rzcannot have k below 1z cannot have k=%d greater than %dN)AssertionError ValueErrorrk)rrr'rrrr szKFold.__init__ccs|j}|j}tt||}t|D]\}tj|td}||dkr^d||||d|<nd|||d<t|}||fVq&dS)Nr T) rr'intr ceilr rrr)rrr'jrrrrrrrs   zKFold.__iter__cCsd|jj|jj|j|jfS)Nz%s.%s(n=%i, k=%i))rrrrr'rrrrrs zKFold.__repr__Nrrrrrr$s"r$c@s(eZdZdZddZddZddZdS) LeaveOneLabelOutzy Leave-One-Label_Out cross-validation iterator: Provides train/test indexes to split data in train test sets cCs ||_dS)a Leave-One-Label_Out cross validation: Provides train/test indexes to split data in train test sets Parameters ---------- labels : list List of labels Examples -------- >>> from scikits.learn import cross_val >>> X = [[1, 2], [3, 4], [5, 6], [7, 8]] >>> y = [1, 2, 1, 2] >>> labels = [1, 1, 2, 2] >>> lol = cross_val.LeaveOneLabelOut(labels) >>> for train_index, test_index in lol: ... print "TRAIN:", train_index, "TEST:", test_index ... X_train, X_test, y_train, y_test = cross_val.split(train_index, test_index, X, y) ... print X_train, X_test, y_train, y_test TRAIN: [False False True True] TEST: [ True True False False] [[5 6] [7 8]] [[1 2] [3 4]] [1 2] [1 2] TRAIN: [ True True False False] TEST: [False False True True] [[1 2] [3 4]] [[5 6] [7 8]] [1 2] [1 2] N)labels)rr-rrrr szLeaveOneLabelOut.__init__ccsVtj|jdd}t|D]6}tjt|td}d|||k<t|}||fVqdS)NT)copyr )r r!r-uniquerlenrr)rr-rrrrrrrs   zLeaveOneLabelOut.__iter__cCsd|jj|jj|jfS)Nz%s.%s(labels=%s))rrrr-rrrrrs zLeaveOneLabelOut.__repr__Nrrrrrr,s" r,cGs@g}|D]2}t|}||}||}||||q|S)zx For each arg return a train and test subsets defined by indexes provided in train_indexes and test_indexes )r Z asanyarrayappend)Z train_indexesZ test_indexesargsretargZ arg_trainZarg_testrrrsplits   r5c@s*eZdZdZd ddZddZd d ZdS) KStepAheadzn KStepAhead cross validation iterator: Provides fit/test indexes to split data in sequential sets r(NTcCs<||_||_|dkr&tt|d}||_||_||_dS)a= KStepAhead cross validation iterator: Provides train/test indexes to split data in train test sets Parameters ---------- n: int Total number of elements k : int number of steps ahead start : int initial size of data for fitting kall : bool if true. all values for up to k-step ahead are included in the test index. If false, then only the k-th step ahead value is returnd Notes ----- I do not think this is really useful, because it can be done with a very simple loop instead. Useful as a plugin, but it could return slices instead for faster array access. Examples -------- >>> from scikits.learn import cross_val >>> X = [[1, 2], [3, 4]] >>> y = [1, 2] >>> loo = cross_val.LeaveOneOut(2) >>> for train_index, test_index in loo: ... print "TRAIN:", train_index, "TEST:", test_index ... X_train, X_test, y_train, y_test = cross_val.split(train_index, test_index, X, y) ... print X_train, X_test, y_train, y_test TRAIN: [False True] TEST: [ True False] [[3 4]] [[1 2]] [2] [1] TRAIN: [ True False] TEST: [False True] [[1 2]] [[3 4]] [1] [2] Ng?)rr'r)r truncstartkall return_slice)rrr'r8r9r:rrrr s'zKStepAhead.__init__c cs|j}|j}|j}|jrpt|||D]F}td|d}|jrLt|||}nt||d||}||fVq&nvt|||D]f}tj|t d}d|d|<tj|t d}|jrd||||<nd|||d||<||fVq~dS)Nr(r T) rr'r8r:r slicer9r rr) rrr'r8rZ train_sliceZ test_slicerrrrrrPs$  zKStepAhead.__iter__cCsd|jj|jj|jfSrrrrrrrkszKStepAhead.__repr__)r(NTTrrrrrr6s 0r6) rZstatsmodels.compat.pythonrZnumpyr itertoolsrrrr$r,r5r6rrrrs  4;A: