A New Splitting Criterion for Better Interpretable Trees
A new splitting criterion for classification trees that generates better decision rules in terms of interpretability is proposed in this paper. The criterion is designed to find homogeneous rules that describe a significant number of instances with a short length. The proposed criterion considers on...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2020-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/9054987/ |
| _version_ | 1818330219783651328 |
|---|---|
| author | Sangheum Hwang Hyeon Gyu Yeo Jung-Sik Hong |
| author_facet | Sangheum Hwang Hyeon Gyu Yeo Jung-Sik Hong |
| author_sort | Sangheum Hwang |
| collection | DOAJ |
| description | A new splitting criterion for classification trees that generates better decision rules in terms of interpretability is proposed in this paper. The criterion is designed to find homogeneous rules that describe a significant number of instances with a short length. The proposed criterion considers only one side of a split to generate highly homogeneous rules and concurrently utilizes a function of sample ratios with an adjustable hyperparameter to control the coverage of rules. The distinctive feature of the proposed method is that it is applied adaptively at every split. We also introduce an efficient heuristic algorithm to determine an appropriate hyperparameter value for every split. Experimental results evaluated over 17 benchmark datasets show that the proposed criterion combined with the proposed heuristic constructs a better interpretable decision tree. It is verified through quantitative and qualitative analysis that the constructed tree produces highly interpretable rules, and its predictive performance is comparable to that of other popular criteria. |
| first_indexed | 2024-12-13T13:00:29Z |
| format | Article |
| id | doaj.art-660c005d4c464fc7872d00e6c802beb5 |
| institution | Directory Open Access Journal |
| issn | 2169-3536 |
| language | English |
| last_indexed | 2024-12-13T13:00:29Z |
| publishDate | 2020-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj.art-660c005d4c464fc7872d00e6c802beb52022-12-21T23:45:03ZengIEEEIEEE Access2169-35362020-01-018627626277410.1109/ACCESS.2020.29852559054987A New Splitting Criterion for Better Interpretable TreesSangheum Hwang0https://orcid.org/0000-0003-2136-296XHyeon Gyu Yeo1https://orcid.org/0000-0001-8597-1399Jung-Sik Hong2https://orcid.org/0000-0001-5579-0968Department of Industrial and Information Systems Engineering, Seoul National University of Science and Technology, Seoul, South KoreaDepartment of Industrial and Information Systems Engineering, Seoul National University of Science and Technology, Seoul, South KoreaDepartment of Industrial and Information Systems Engineering, Seoul National University of Science and Technology, Seoul, South KoreaA new splitting criterion for classification trees that generates better decision rules in terms of interpretability is proposed in this paper. The criterion is designed to find homogeneous rules that describe a significant number of instances with a short length. The proposed criterion considers only one side of a split to generate highly homogeneous rules and concurrently utilizes a function of sample ratios with an adjustable hyperparameter to control the coverage of rules. The distinctive feature of the proposed method is that it is applied adaptively at every split. We also introduce an efficient heuristic algorithm to determine an appropriate hyperparameter value for every split. Experimental results evaluated over 17 benchmark datasets show that the proposed criterion combined with the proposed heuristic constructs a better interpretable decision tree. It is verified through quantitative and qualitative analysis that the constructed tree produces highly interpretable rules, and its predictive performance is comparable to that of other popular criteria.https://ieeexplore.ieee.org/document/9054987/Data miningdecision treesinterpretabilitysplitting criterion |
| spellingShingle | Sangheum Hwang Hyeon Gyu Yeo Jung-Sik Hong A New Splitting Criterion for Better Interpretable Trees IEEE Access Data mining decision trees interpretability splitting criterion |
| title | A New Splitting Criterion for Better Interpretable Trees |
| title_full | A New Splitting Criterion for Better Interpretable Trees |
| title_fullStr | A New Splitting Criterion for Better Interpretable Trees |
| title_full_unstemmed | A New Splitting Criterion for Better Interpretable Trees |
| title_short | A New Splitting Criterion for Better Interpretable Trees |
| title_sort | new splitting criterion for better interpretable trees |
| topic | Data mining decision trees interpretability splitting criterion |
| url | https://ieeexplore.ieee.org/document/9054987/ |
| work_keys_str_mv | AT sangheumhwang anewsplittingcriterionforbetterinterpretabletrees AT hyeongyuyeo anewsplittingcriterionforbetterinterpretabletrees AT jungsikhong anewsplittingcriterionforbetterinterpretabletrees AT sangheumhwang newsplittingcriterionforbetterinterpretabletrees AT hyeongyuyeo newsplittingcriterionforbetterinterpretabletrees AT jungsikhong newsplittingcriterionforbetterinterpretabletrees |