precise_example: Example Data Sets

In topepo/tidyposterior: Bayesian Analysis to Compare Models using Resampling Statistics

Example Data Sets

Description

Example Data Sets

Details

Several data sets are contained in the package as examples. Each simulates an rset object but the splits columns are not included to save space.

• precise_example contains the results of the classification analysis of a real data set using 10-fold CV. The holdout data sets contained thousands of examples and have precise performance estimates. Three models were fit to the original data and several performance metrics are included.

• noisy_example was also generated from a regression data simulation. The original data set was small (50 samples) and 10-repeated of 10-fold CV were used with four models. There is an excessive of variability in the results (probably more than the resample-to-resample variability). The RMSE distributions show fairly right-skewed distributions.

• concrete_example contains the results of the regression case study from the book Applied Predictive Modeling. The original data set contained 745 samples in the training set. 10-repeats of 10-fold CV was also used and 13 models were fit to the data.

• ts_example is from a data set where rolling-origin forecast resampling was used. Each assessment set is the summary of 14 observations (i.e. 2 weeks). The analysis set consisted of a base of about 5,500 samples plus the previous assessment sets. Four regression models were applied to these data.

• ex_object objects were generated from the two_class_dat data in the modeldata package. Basic 10-fold cross validation was used to evaluate the models. The posterior_samples object is samples of the posterior distribution of the model ROC values while contrast_samples are posterior probabilities form the differences in ROC values.

Value

Tibbles with the additional class rset

Examples

data(precise_example)
precise_example

topepo/tidyposterior documentation built on Oct. 18, 2023, 8:30 p.m.