• Use the Random Data widget to create a data set with 100 data instances, and 500 variables with normal random distribution, and one Bernouli variable (p = 0.5).
  • Use Select Columns to designate the Bernouli variable as target.
  • Use Test and Score to evaluate performance of classification tree on this data, using
    • testing on training data
    • cross validation

For classification tree, set the pruning parameters to defaults (2, 5, 100, 95), or even disable them.

Generate the random data a few times (by clicking "Generate" button at the bottom) to get a general impresion about the results in both cases.

Zadnja sprememba: ponedeljek, 7. marec 2022, 16.27