forked from paobranco/UBL
-
Notifications
You must be signed in to change notification settings - Fork 0
/
CHANGES
37 lines (32 loc) · 1.96 KB
/
CHANGES
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
VERSION 0.0.7 (2021-03-29)
- added MetaCost methods for classification and regression (based on MeatCost from Domingos, 1999)
- added methods for introducing a Neighbourhood Bias in smoteR and under-sampling for regression and classification
- added SMOGN algorithm for imbalanced regression
- added MetaUtilRegress algorithm
- added bagging method for dealing with imbalanced regression problems
- improvements of functions documentation
- bugs correction in HVDM function (thanks to Miriam Santos and Arne Camps for noticing this)
- correction added in ADASYN function to solve examples generation when there are no nearest neighbours in the baseClass
VERSION 0.0.6 (2017-06-20)
- added adasyn algorithm for classification
- added method for interpolating utility/benefit/cost surfaces
- added utility-based metrics for classification and regression tasks (utility, cost and benefit)
- added a function for utility-based learning optimization for classification (from Elkan 2001)
- added a function for utility-based learning optimization for regression
- added a function that returns the distances between all pairs of
example computed according to a selected distance metric
- bugs correction in SmoteRegress function (caused by constant features)
- bugs correction on RandOverClassif (caused by classes with only one example)
- bugs correction on neighbours function (Fortran code)
- routines registration added
VERSION 0.0.5 (2016-07-13)
- added two synthetic data sets, for classification and regression
- correction of bugs
- added verifications of percentages provided for under/over-sampling that should be less/greater than 1 in each case.
- improvements of functions documentation
- improved warnings issued when no change is performed in the data set
- vignettes restructured to use the synthetic data sets introduced in this version
VERSION 0.0.4 (2016-04-28)
- correction of bugs related with FORTRAN code
VERSION 0.0.3 (2016-04-15)
- first version