NEWS.md

In Marchen/cv.models: Model Selection and Hyper-Parameter Tuning by Cross Validation

2019-07-09: 0.1.4

• Add Spearman's rho and Kendall's tau as performance measures for regression models.
• Start CI with Travis-CI.
• Bug fixes.

2018-11-10: 0.1.3

• Add group argument which controls user defined grouping for cross validation.

2018-09-14: 0.1.2

• Support model adapter > 0.1.0.
• Bug fix: occasionally fail to detect number of CPU cores.
• Add English vignette.

2018-01-31: 0.1.1

• Bug fix: cross validation is not conducted when model object or call produced by call() function is specified.

2018-01-28: 0.1.0

• Now find.best.models() uses the seed of random numbers in cv.models object if seed in the object is not NULL.
• Replaced dependency on pROC with OptimalCutpoints.
• Add calculation of Cohen's Kappa for metrics of classification models.
• Bug fix: specifying 'seed' returns same result even with cluster.
• Minor bug fix of print.cv.best.models().

2017-09-10: 0.0.4

• Experimental: implemented log-likelihood and likelihood based R squared values for classification models.
• Bug fix: wrong assignment of CPU cores.

2017-07-07: 0.0.3

• Bug fix: occasionally evaluation of model fails.

2017-07-07: 0.0.2

• Bug fix: correspondence between the response variable and the predicted values was broken.

2017-07-05: 0.0.1

• Started versioning.

Marchen/cv.models documentation built on Sept. 2, 2020, 2:04 a.m.