<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE ArticleSet PUBLIC "-//NLM//DTD PubMed 2.7//EN" "https://dtd.nlm.nih.gov/ncbi/pubmed/in/PubMed.dtd">
<ArticleSet>
<Article>
<Journal>
				<PublisherName>University of Sistan and Baluchestan</PublisherName>
				<JournalTitle>Iranian Journal of Fuzzy Systems</JournalTitle>
				<Issn>1735-0654</Issn>
				<Volume></Volume>
				<Issue>Articles in Press</Issue>
				<PubDate PubStatus="epublish">
					<Year>2026</Year>
					<Month>10</Month>
					<Day>31</Day>
				</PubDate>
			</Journal>
<ArticleTitle>Faithful regularised classification over linguistic fuzzy rule activations</ArticleTitle>
<VernacularTitle></VernacularTitle>
			<FirstPage>1</FirstPage>
			<LastPage>21</LastPage>
			<ELocationID EIdType="pii">10290</ELocationID>
			
<ELocationID EIdType="doi">10.22111/ijfs.2026.56189.10009</ELocationID>
			
			<Language>EN</Language>
<AuthorList>
<Author>
					<FirstName>S.</FirstName>
					<LastName>Atmakuri Davidsen</LastName>
<Affiliation>Hamisoke Labs, R&amp;D</Affiliation>

</Author>
<Author>
					<FirstName>M.</FirstName>
					<LastName>Padmavathamma</LastName>
<Affiliation>S. V. University, Dept of Computer Science</Affiliation>

</Author>
</AuthorList>
				<PublicationType>Journal Article</PublicationType>
			<History>
				<PubDate PubStatus="received">
					<Year>2026</Year>
					<Month>07</Month>
					<Day>20</Day>
				</PubDate>
			</History>
		<Abstract>Achieving competitive accuracy while exposing exact prediction computations remains difficult for tabular classifiers. We  introduce and evaluate FuzzyRuleSVM, a regularised classifier that maps each instance into linguistic fuzzy antecedent  activations and trains a linear SVM in that bounded feature space. Unlike classical fuzzy SVMs, which fuzzify sample  weights in the objective, FuzzyRuleSVM substitutes the feature space: every dimension is a named antecedent, and its  coefficient supplies the signed class effect. The decision function therefore decomposes exactly into named rule- activation contributions. This is algebraic fidelity, not cognitive interpretability. On twenty-one benchmark and domain- specific datasets, including Steel Plates Faults as an industrial manufacturing case, FuzzyRuleSVM achieves competitive  balanced accuracy while exposing exact rule-contribution decompositions, runtime profiles, and compactness/stability  metrics. The contribution is faithful regularised classification over linguistic rule activations. </Abstract>
		<ObjectList>
			<Object Type="keyword">
			<Param Name="value">Fuzzy rule classification</Param>
			</Object>
			<Object Type="keyword">
			<Param Name="value">Support vector machine</Param>
			</Object>
			<Object Type="keyword">
			<Param Name="value">regularized classification</Param>
			</Object>
			<Object Type="keyword">
			<Param Name="value">linguistic rule activation</Param>
			</Object>
			<Object Type="keyword">
			<Param Name="value">algebraic explanation</Param>
			</Object>
		</ObjectList>
<ArchiveCopySource DocType="pdf">https://ijfs.usb.ac.ir/article_10290_8af345704e442a54eb3649a78c928d5e.pdf</ArchiveCopySource>
</Article>
</ArticleSet>
