|نویسندگان||Mohsen Rahmanian-Eghbal Mansoori-Mohammad Taheri|
|همایش||10th International Conference on Computer and Knowledge Engineering|
|تاریخ برگزاری همایش||2020-10-29|
|محل برگزاری همایش||مشهد|
|سطح همایش||بین المللی|
The data readability,complexity reduction of learning algorithmsand increase predictabilityare the most important reasonsfor using feature selection methods, especially when there exist lotsof features.In recent years, unsupervised feature selection techniques are well explored. In this paper, we proposed an unsupervised feature selection algorithm using multivariate-symmetrical-uncertainty based feature clustering, Feature Selection-based Virtual Feature Representative (FSVFR).The main idea of FSVFR isas follows: First, it selectsthe clustercentersbased on the similarity density of the neighbors of each feature;afterassigning the featuresto the clusters, the virtual representativeisgeneratedin such a way that contains maximum common information withcluster’s membersand minimumsimilaritywithother representatives.Thesesteps continues until there is no more change in the representatives.Second, a feature that has the most common information in each cluster is selected as its representative. The experimental resultson benchmark datasets demonstrate the effectiveness ofour approaches as comparedto the twocommonmethods.