9 Empirical results - adaa-polsl/RuleKit GitHub Wiki

RuleKit 1

The experiments described in this section were included in the RuleKit paper:

Gudyś, A, Sikora, M, Wróbel, Ł (2020) RuleKit: A comprehensive suite for rule-based learning, Knowledge-Based Systems, 194: 105480

The tests were performed in stratified 10-fold cross validation scheme on 50 classification and 47 regression datasets. Following software packages were evalulated.

Classification:

  • RuleKit,
  • RIPPER,
  • CN2,
  • AQ,

Regression:

  • RuleKit,
  • M5Rules.

All algorithms were run with default parameters. Detailed results of all investigated algorithms on particular datasets can be found here.

RuleKit 2

The experiments described in this section were included in the RuleKit 2 paper.

The tests were performed in stratified 10-fold cross validation scheme on 50 classification, 48 regression, and 22 survival datasets. The results can be found here.