Numerical Data Classification via Distance-Based Similarity Measures of Fuzzy Parameterized Fuzzy Soft Matrices

Memis S., ENGİNOĞLU S., ERKAN U.

IEEE ACCESS, vol.9, pp.88583-88601, 2021 (SCI-Expanded) identifier

  • Publication Type: Article / Article
  • Volume: 9
  • Publication Date: 2021
  • Doi Number: 10.1109/access.2021.3089849
  • Journal Name: IEEE ACCESS
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Compendex, INSPEC, Directory of Open Access Journals
  • Page Numbers: pp.88583-88601
  • Keywords: Fuzzy sets, soft sets, fpfs-matrices, similarity measure, classification, supervised learning, SET-THEORY
  • Çanakkale Onsekiz Mart University Affiliated: Yes

Abstract

In this paper, we first define eight pseudo-metrics and eight pseudo-similarities based on these pseudo-metrics overfpfs-matrices. We then propose a new classification algorithm, i.e. Fuzzy Parameterized Fuzzy Soft Euclidean Classifier (FPFS-EC), based on Euclidean pseudo-similarity. After that, we compare FPFS-EC with Support Vector Machines (SVM), Fuzzy k-Nearest Neighbor (Fuzzy kNN), Fuzzy Soft Set Classifier (FSSC), FussCyier, Fuzzy Soft Set Classification Using Hamming Distance (HDFSSC), and Fuzzy kNN Based on the Bonferroni Mean (BM-Fuzzy kNN) in terms of the performance criteria - namely accuracy, precision, recall, macro F-score, and micro F-score - and running time by using 18 real-world datasets in the UCI machine learning repository. The results show that FPFS-EC performs better in the occurrence of the 13 of 18 datasets in question than SVM, Fuzzy kNN, FSSC, FussCyier, HDFSSC, and BM-Fuzzy kNN.