作者: Umit Kilic , Mumine Kaya Keles
DOI: 10.1109/ASYU.2018.8554004
关键词:
摘要: Relevant and irrelevant features compose data. Evaluation of these is the fundamental task for classification clustering processes during this processes, induce obtaining false results. Likewise, due to relevant features' direct effect on results can be more correct stable. This also represents aim feature selection process that tries achieve as high possible with small subset. In study, Artificial Bee Colony (ABC) algorithm based method updated employed Z-Alizadeh Sani data set consists 56 including class attribute collected from 303 patients. 16 are selected by ABC method. Also, accuracy F-measure values measured 89.4% 0.894 respectively, which higher than produced raw dataset.