This repositorie uses implementation code for a sample size estimation task. The source code for the problem of the sample size estimation for linear models.
This repository contains computation experiment to the article: Numerical methods of minimum sufficient sample size estimation for linear and logistic regression.
This part of computation experiment shows how different models depends on available sample size. We are using Boston Housing dataset for illustration. The dataset is available through the link.
This part of computation experiment shows how different models works on different datasets. We are using different public available datasets. A description of all datasets is available through the link.
| Boston m* | Diabetes m* | ForestFires m* | Servo m* | NBA m* | |
| LagrangeEstimator | 18 | 25 | 44 | 38 | 218 |
| LikelihoodRatioEstimator | 17 | 25 | 43 | 18 | 110 |
| WaldEstimator | 66 | 51 | 46 | 76 | 200 |
| BootstrapEstimator | 178 | 441 | 171 | 120 | 1328 |
| CrossValidationEstimator | 113 | 117 | 86 | 60 | 405 |
| APVCEstimator | 98 | 167 | 351 | 20 | 1328 |
| ACCEstimator | 228 | 441 | 346 | 65 | 1328 |
| ALCEstimator | 98 | 267 | 516 | 25 | 1328 |
| MaxUtilityEstimator | 148 | 172 | 206 | 105 | 925 |
