Large-Scale Fuzzy Least Squares Twin SVMs for Class Imbalance Learning
Authors: M. A. Ganaie, M. Tanveer and C. -T. Lin
Publication: IEEE Transactions on Fuzzy Systems (TFS)
Issue: Volume 30, Issue 11 – November 2022
Pages: 4815-4827

Abstract: Twin support vector machines (TSVMs) have been successfully employed for binary classification problems. With the advent of machine learning algorithms, data have proliferated and there is a need to handle or process large-scale data. TSVMs are not successful in handling large-scale data due to the following: 1) the optimization problem solved in the TSVM needs to calculate large matrix inverses, which makes it an ineffective choice for large-scale problems; 2) the empirical risk minimization principle is employed in the TSVM and, hence, may suffer due to overfitting; and 3) the Wolfe dual of TSVM formulation involves positive-semidefinite matrices, and hence, singularity issues need to be resolved manually. Keeping in view the aforementioned shortcomings, in this article, we propose a novel large-scale fuzzy least squares TSVM for class imbalance learning (LS-FLSTSVM-CIL). We formulate the LS-FLSTSVM-CIL such that the proposed optimization problem ensures that: 1) no matrix inversion is involved in the proposed LS-FLSTSVM-CIL formulation, which makes it an efficient choice for large-scale problems; 2) the structural risk minimization principle is implemented, which avoids the issues of overfitting and results in better performance; and 3) the Wolfe dual formulation of the proposed LS-FLSTSVM-CIL model involves positive-definite matrices. In addition, to resolve the issues of class imbalance, we assign fuzzy weights in the proposed LS-FLSTSVM-CIL to avoid bias in dominating the samples of class imbalance problems. To make it more feasible for large-scale problems, we use an iterative procedure known as the sequential minimization principle to solve the objective function of the proposed LS-FLSTSVM-CIL model. From the experimental results, one can see that the proposed LS-FLSTSVM-CIL demonstrates superior performance in comparison to baseline classifiers. To demonstrate the feasibility of the proposed LS-FLSTSVM-CIL on large-scale classification problems, we evaluate the classification models on the large-scale normally distributed clustered (NDC) dataset. To demonstrate the practical applications of the proposed LS-FLSTSVM-CIL model, we evaluate it for the diagnosis of Alzheimer’s disease and breast cancer disease. Evaluation on NDC datasets shows that the proposed LS-FLSTSVM-CIL has feasibility in large-scale problems as it is fast in comparison to the baseline classifiers.

Index Terms: Class imbalance, machine learning, maximum margin, support vector machines (SVMs), twin support vector machine (TSVM).
IEEE Xplore Link: https://ieeexplore.ieee.org/document/9740453

Professional Editing Services

Sometimes TFS receives submissions contain poor English usage and readability. Such manuscripts often get rejected because of extremely poor readability. Authors, at their own cost, may take the help of American Journal Experts for pre-submission professional editing services. An author willing to get assistance with English grammar and usage prior to submitting his/her manuscripts for review or during the review process can go directly to www.aje.com/go/ieee/ to submit a manuscript for copy editing. Various levels of editing services are available. Cost estimates as well as required time are available immediately on line. We cannot guarantee that the linguistic quality of an edited manuscript will meet your expectation. As expected, an edited manuscript will undergo usual reviews.

IEEE Transactions on Fuzzy Systems publishes Special Issues on emerging topics guest edited by distinguished researchers in fuzzy systems and related areas. 

On-going Special Issue(s):

• IEEE TFS Special Issue on Deep Neuro-Fuzzy Approaches for Intelligent Big Data Processing (Extended to 30 March 2024)

Please see below information about how a special issue is organized.

Phase 1 - Proposal:
Interested researchers should submit a proposal for the Special Issue on an emergent topic in fuzzy systems or its closely related areas. The proposal should be emailed to the Editor-in-Chief – Prof. Dongrui Wu, School of Artificial Intelligence and Automation, Huazhong University of Science and Technology) drwu09@gmail.com

The proposal should include at least the following components:

• the theme of the Special Issue
• its relevance/importance and need in the present context
• a list of specific topics focusing the Special Issue
• a list of potential authors (to ensure that there are enough researchers to support such a special issue)
• short bios of the proposed Guest Editors including personal web page links
• a feasible timetable.

The Special Issue will be organised based on an open call for papers - we do not consider Special Issues based on papers from a conference.

Phase 2-Evaluation:
The Editor in Chief will evaluate the proposal. The evaluation takes into account various issues including the technical merit, need and relevance, timeliness, and feasibility of such a Special Issue. The Editor in Chief makes the final decision on the Special Issue proposal. The decision could be acceptance, rejection or a revision of the proposal.

Phase 3- Call for Papers:
If a proposal is accepted, the Guest Editor is asked to prepare a call for papers (CFP) so that it can be published in our transactions. The CFP should include all relevant information such as the theme, topic, deadlines, and submission guidelines.
The CFP must include an instruction to authors suggesting them to ensure the ‘Special Edition’ option is ticked on submission and mention something like "This paper is for the special issue on XXXX" in their covering letter to the Editor at the time of submission through Manuscript-Central.
The CFP is usually announced in the journal’s web site, circulated through the TFS email and is published in the CIS Transactions and Magazine.

Phase 4- Processing:
Papers submitted for a Special Issue are assigned to the Guest Editor(s), acting as Associate Editors. Guest Editors(s) are responsible for inviting reviewers, chasing overdue reviews and handling the review process (help is available from the Journal Administrator). The review process for Special Issue papers is exactly the same as that for regular transactions papers. If a Guest Editor is an author of a paper submitted for the Special Issue, then reviewing of that manuscript is handled by a different Associate Editor chosen by the EiC. All decisions are subject to final approval by the Editor-in-Chief.
A Guest Editor can be nominated to make preliminary checks on all submitted papers. All Guest Editors must then be jointly involved – the papers will be distributed as equally as possible.
We will provide an up-to-date spreadsheet of the status of all papers as requested.

Phase 5-Publication:
Once the special issue is complete, the Editor in Chief requests the Guest Editors to write an Editorial to the Special Issue (usually not more than 2 formatted transactions pages) for inclusion in the Special Issue. The Special Issue is then published as soon as possible, once a date has been agreed with our Journals Production Manager.

General Points:
We do not encourage Guest Editors to submit papers in the Special Issue and under no circumstances should any of the Guest Editors submit more than ONE manuscript for the Special Issue.
Although multiple Guest Editors for a special issue are allowed, it is better to keep the number of Guest Editors as low as possible.
This is just a broad guideline and there may be other important points not listed here. Please email drwu09@gmail.com with any queries.

​

Information for Reviewers

We endeavor to administer a fair and timely review process for all of our manuscripts and to publish only papers of the highest quality - this is largely dependent upon the efforts of our reviewers.

Conduct of the Reviewer

To guarantee fairness to the author, the reviewer of a manuscript should abide by a number of guidelines, including, but not limited to:

• Respond promptly within the allotted time
• Provide sound, constructive reviews
• Assume that manuscripts submitted for publication are not meant to be public
• Do not use material from a manuscript you have reviewed
• Do not share material from a manuscript you have reviewed with others

· Do not distribute copies of a manuscript you have been asked to review unless the material is already public

• Tell the Associate Editor, Guest Editor, and Editor-in- Chief if there are any conflicts of interest involved in reviewing a manuscript

IEEE Review Process Statement (8.2.2.A Review Process)

IEEE requires that referees treat the contents of papers under review as confidential information not to be disclosed to others before publication. It is expected that no one with access to a paper under review will make any inappropriate use of the special knowledge that access provides.

To access Section 8.2.2.A – The Review Process in its entirety, visit the IEEE Publication Services and Products Board (PSPB) Operations Manual .

Length of Review Process

The length of the review process varies between papers, and many factors affect the time it takes to review a paper. The time it takes for an Associate Editor to be assigned, and for qualified referees to be located who are available to handle the review of a paper, depends on the complexity and the type of material submitted for review.

We make a concerted effort to keep the review time to a minimum. Each submission placed into review is sent to at least three reviewers. The review process may take approximately six months to be completed.

Preliminary or Conference Version

If the authors submit a previously published conference paper, please check the manuscript to determine whether a sufficient amount of new material has been added to warrant publication. Section 8.2.1.B.9 of the IEEE PSPB Operations Manual specifies that authors should only submit original work that has neither appeared elsewhere for publication, nor which is under review for another refereed publication. If an author used their own previously published material as a basis for a new submission, then they are required to cite the previous work(s) and clearly indicate how the new submission offers substantively novel or different contributions beyond those of the previously published work(s).

Also included should be a brief description of the differences between the papers.

Comments Paper

A comments paper is a paper commenting on an error one has found in, or a disagreement one has with, a previously published paper.

When submitting a comments paper, the author must include the title and/or number of the previously published manuscript. The journal coordinator will retrieve as much information as possible about the paper being commented on and will forward the materials to the Editor-in-Chief (EIC). The EC will assign the comments paper to the Associate Editor (AE) who was assigned to the previously published paper. If that AE is no longer available, the EIC will assign the comments paper to another AE whose expertise areas closely match that paper's topic.

Supplemental Material

IEEE journals accept supplemental materials for review with regular paper submissions. Types of supplemental material can include: proofs, code, experimental data, appendices, animations and audio files. These materials may be published on our digital library with the electronic version of the paper, where they can be accessed for free. A pointer to the supplemental material will be included in the printed version.

Supplemental material must also undergo peer review, so questions pertaining to the supplemental material (if applicable) are included in the review form.

Recommended Reading

It is recommended that all reviewers read the paper titled Task of the Referee (PDF), about reviewing manuscripts.

IEEE Links

http://ieeeauthorcenter.ieee.org/publish-with-ieee/peer-review-ieee/what-is-peer-review/

https://supportcenter.ieee.org/app/answers/detail/a_id/1445/~/reviewer---instructions-for-manuscript-central

http://ieeeauthorcenter.ieee.org/publish-with-ieee/peer-review-ieee/ethics-peer-review/