Intrusion Detection Method Based on Multiscale Triplet Temporal Windows and Hierarchical Attention

Intrusion Detection Method Based on Multiscale Triplet Temporal Windows and Hierarchical Attention | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Intrusion Detection Method Based on Multiscale Triplet Temporal Windows and Hierarchical Attention Hui Zhang This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9453801/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 7 You are reading this latest preprint version Abstract In order to address the complex temporal characteristics of IoT/IIoT network traffic, which include short-term bursts, phased accumulation, and long-range evolution, the papers propose an intrusion detection method combining a three-scale temporal window with hierarchical attention. This approach models information across short, medium, and long time windows in parallel, highlighting critical time segments within windows through temporal attention and adaptively allocating contributions across different time scales between windows via scale attention. It further mitigates class imbalance issues by incorporating a weighted loss function. The experimental results on multiple public datasets including Bot IoT, TON IoT, N-BaIoT, and MQTTset show that the paper's proposed method achieves the highest accuracy of 98.83% and F1 score of 98.69% in binary classification tasks; In multi class scenarios, Macro-F1 shows an average improvement of approximately 2.1% -2.7% compared to strong baseline methods. The results verified the effectiveness of multi-scale modeling and hierarchical attention mechanism in improving intrusion detection accuracy and minority class recognition ability. intrusion detection Multi scale temporal window Hierarchical attention IoT/IIoT network Unbalanced learning Full Text Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Review Version 1 posted Reviews received at journal 10 May, 2026 Reviewers agreed at journal 04 May, 2026 Reviewers invited by journal 29 Apr, 2026 Editor invited by journal 28 Apr, 2026 Editor assigned by journal 22 Apr, 2026 Submission checks completed at journal 22 Apr, 2026 First submitted to journal 17 Apr, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9453801","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":634032874,"identity":"fe1f5e05-d0d1-44a8-be87-8a2d07f1267c","order_by":0,"name":"Hui Zhang","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA+klEQVRIie3QsWrDMBCA4RMCd7ngVcJOnkEloA4d/CqGQrx4KwSPAcNNoV3dF8msYPCUtmtLlmTpnAwNDoQQpaVLwY7HQvUPQoj7hhOAy/UHiwp7MAL0/cf1yt5DAK+dsB8iiwVXAAo7E1Am9UQnwoP8Y/1JZQjmucq2dED/imKos1kj8cLqZtg/lsjyh9H7EymU08qw6WLZSFDEOpBUIgejlz1L1Fsy4YyaiRDJ7ot4EOv7TkSJVMutJQip5t9kZC6RccAoQSGqO1m8DM+7xPO2XaIimck93UbRaz7fbMaDgf2x61WdNZNzHH+/mNZ5G6svTbhcLtf/7gRMxFQHYoU+EwAAAABJRU5ErkJggg==","orcid":"","institution":"Yancheng Teachers University","correspondingAuthor":true,"prefix":"","firstName":"Hui","middleName":"","lastName":"Zhang","suffix":""}],"badges":[],"createdAt":"2026-04-18 03:38:46","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9453801/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9453801/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":108807203,"identity":"76638af6-a0c3-42da-a0ff-b4b042123d61","added_by":"auto","created_at":"2026-05-08 15:30:18","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1182846,"visible":true,"origin":"","legend":"","description":"","filename":"article.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9453801/v1_covered_3ddc1fa1-c618-4abb-a876-0fd7c1dfb131.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Intrusion Detection Method Based on Multiscale Triplet Temporal Windows and Hierarchical Attention","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":true,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":true,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"discover-computing","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"Learn more about [Discover Computing](https://link.springer.com/journal/10791)","snPcode":"10791","submissionUrl":"https://submission.springernature.com/new-submission/10791/3","title":"Discover Computing","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Discover Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"intrusion detection, Multi scale temporal window, Hierarchical attention, IoT/IIoT network, Unbalanced learning","lastPublishedDoi":"10.21203/rs.3.rs-9453801/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9453801/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eIn order to address the complex temporal characteristics of IoT/IIoT network traffic, which include short-term bursts, phased accumulation, and long-range evolution, the papers propose an intrusion detection method combining a three-scale temporal window with hierarchical attention. This approach models information across short, medium, and long time windows in parallel, highlighting critical time segments within windows through temporal attention and adaptively allocating contributions across different time scales between windows via scale attention. It further mitigates class imbalance issues by incorporating a weighted loss function. The experimental results on multiple public datasets including Bot IoT, TON IoT, N-BaIoT, and MQTTset show that the paper's proposed method achieves the highest accuracy of 98.83% and F1 score of 98.69% in binary classification tasks; In multi class scenarios, Macro-F1 shows an average improvement of approximately 2.1% -2.7% compared to strong baseline methods. The results verified the effectiveness of multi-scale modeling and hierarchical attention mechanism in improving intrusion detection accuracy and minority class recognition ability.\u003c/p\u003e","manuscriptTitle":"Intrusion Detection Method Based on Multiscale Triplet Temporal Windows and Hierarchical Attention","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-05-07 19:29:37","doi":"10.21203/rs.3.rs-9453801/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"editorInvitedReview","content":"","date":"2026-05-10T10:05:22+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"335396777040941377046195562910792045851","date":"2026-05-04T09:45:36+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-04-29T09:04:52+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2026-04-28T08:45:31+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-04-22T14:09:29+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-04-22T14:08:37+00:00","index":"","fulltext":""},{"type":"submitted","content":"Discover Computing","date":"2026-04-18T03:28:52+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"discover-computing","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"Learn more about [Discover Computing](https://link.springer.com/journal/10791)","snPcode":"10791","submissionUrl":"https://submission.springernature.com/new-submission/10791/3","title":"Discover Computing","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Discover Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"6f5d7503-e08a-4006-b2a2-df4a4b525a05","owner":[],"postedDate":"May 7th, 2026","published":true,"recentEditorialEvents":[{"type":"editorInvitedReview","content":"","date":"2026-05-10T10:05:22+00:00","index":23,"fulltext":""},{"type":"reviewerAgreed","content":"335396777040941377046195562910792045851","date":"2026-05-04T09:45:36+00:00","index":22,"fulltext":""}],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-05-07T19:29:37+00:00","versionOfRecord":[],"versionCreatedAt":"2026-05-07 19:29:37","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9453801","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9453801","identity":"rs-9453801","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}