Clustering by twin support vector machine and least square twin support vector classifier with uniform output coding

2021 ◽

pp. 8-13

Author(s):

Thế Cường Nguyễn ◽

Thanh Vi Nguyen

Keyword(s):

Support Vector Machine ◽

Binary Classification ◽

Least Square ◽

Twin Support Vector Machine ◽

Support Vector ◽

Classification Problems ◽

Training Time ◽

Data Points

In binary classification problems, two classes of data seem to be different from each other. It is expected to bemore complicated due to the number of data points of clusters in each class also be different. Traditional algorithmsas Support Vector Machine (SVM), Twin Support Vector Machine (TSVM), or Least Square Twin Support VectorMachine (LSTSVM) cannot sufficiently exploit information about the number of data points in each cluster of the data.Which may be effect to the accuracy of classification problems. In this paper, we propose a new Improvement LeastSquare - Support Vector Machine (called ILS-SVM) for binary classification problems with a class-vs-clusters strategy.Experimental results show that the ILS-SVM training time is faster than that of TSVM, and the ILS-SVM accuracy isbetter than LSTSVM and TSVM in most cases.

Download Full-text

Phishing Website Classification using Least Square Twin Support Vector Machine

International Journal of Innovative Technology and Exploring Engineering - Special Issue ◽

10.35940/ijitee.a3905.119119 ◽

2019 ◽

Vol 9 (1) ◽

pp. 2063-2068

Keyword(s):

Support Vector Machine ◽

Cyber Security ◽

Classification Accuracy ◽

Personal Information ◽

Classification Problem ◽

Least Square ◽

Twin Support Vector Machine ◽

Support Vector ◽

Security Issue

Phishing is one among the luring procedures used by phishing attackers in the means to abuse the personal details of clients. Phishing is earnest cyber security issue that includes facsimileing legitimate website to apostatize online users so as to purloin their personal information. Phishing can be viewed as special type of classification problem where the classifier is built from substantial number of website's features. It is required to identify the best features for improving classifiers accuracy. This study, highlights on the important features of websites that are used to classify the phishing website and form the legitimate ones by presenting a scheme Decision Tree Least Square Twin Support Vector Machine (DT-LST-SVM) for the classification of phishing website. UCI public domain benchmark website phishing dataset was used to conduct the experiment on the proposed classifier with different kernel function and calculate the classification accuracy of the classifiers. Computational results show that DT-LST-SVM scheme yield the better classification accuracy with phishing websites classification dataset

Download Full-text

An Improvement Of Least Square - Twin Support Vector Machine

Research and Development on Information and Communication Technology ◽

10.32913/mic-ict-research.v2021.n1.956 ◽

2021 ◽

Vol 2021 (1) ◽

Author(s):

Thanh Vi Nguyen ◽

Thế Cường Nguyễn

Keyword(s):

Support Vector Machine ◽

Binary Classification ◽

Least Square ◽

Experimental Results ◽

Twin Support Vector Machine ◽

Support Vector ◽

Classification Problems ◽

Training Time ◽

Data Points ◽

Better Than

n binary classification problems, two classes of data seem tobe different from each other. It is expected to be more complicated dueto the number of data points of clusters in each class also be different.Traditional algorithms as Support Vector Machine (SVM), Twin Support Vector Machine (TSVM), or Least Square Twin Support VectorMachine (LSTSVM) cannot sufficiently exploit information about thenumber of data points in each cluster of the data. Which may be effectto the accuracy of classification problems. In this paper, we proposes anew Improved Least Square - Support Vector Machine (called ILS-SVM)for binary classification problems with a class-vs-clusters strategy. Experimental results show that the ILS-SVM training time is faster thanthat of TSVM, and the ILS-SVM accuracy is better than LSTSVM andTSVM in most cases.

Download Full-text