Affinity-based elastic net intuitionistic fuzzy twin support vector machines

Intuitionistic fuzzy twin support vector machines (IFTWSVM) combined the concept of intuitionistic fuzzy sets with twin support vector machines (TWSVM) and showed excellent performance in classification. However, the existing intuitionistic fuzzy number schemes based on the single center and the loc...

Full description

Saved in:
Bibliographic Details
Published in:International journal of machine learning and cybernetics 2024-06, Vol.15 (6), p.2439-2455
Main Authors: Li, Zhishen, Zhang, Peiai
Format: Article
Language:English
Subjects:
Affinity
Algorithms
Artificial Intelligence
Classification
Complex Systems
Computational Intelligence
Control
Datasets
Decision making
Engineering
Fuzzy sets
Fuzzy systems
Handwriting
Mechatronics
Neighborhoods
Original Article
Pattern Recognition
Robotics
Set theory
Slack variables
Support vector machines
Systems Biology
Citations: Items that this one cites
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Intuitionistic fuzzy twin support vector machines (IFTWSVM) combined the concept of intuitionistic fuzzy sets with twin support vector machines (TWSVM) and showed excellent performance in classification. However, the existing intuitionistic fuzzy number schemes based on the single center and the local neighborhood of the sample are difficult to accurately reflect the location information of the sample, and the L1-norm penalty of the slack variable is not well defined from the point of view of geometric points. In view of the above deficiencies, we design a noval intuitionistic fuzzy number scheme and adopt elastic net to penalize the slack variables, propose Affinity-Based Elastic Net Intuitionistic Fuzzy Twin Support Vector Machines (AENIFTWSVM). It calculates the affinity of different classes of samples according to the Support Vector Data Description (SVDD) model in the kernel space, and considers the contribution of samples to the two classes, and the obtained affinity can be used to identify noise information. A series of experimental outcomes on benchmark datasets and handwritten digit dataset support that the proposed model outperforms some existing models.
ISSN:1868-8071
1868-808X
DOI:10.1007/s13042-023-02041-y