Evaluation Benchmark Study for XAI Methods in Arabic Sentiment Analysis

Evaluation Benchmark Study for XAI Methods in Arabic Sentiment Analysis | 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 Evaluation Benchmark Study for XAI Methods in Arabic Sentiment Analysis Youssef Chafiqui, Houda Anoun This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8627067/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Explainable Artificial Intelligence (XAI) is essential for interpreting transformer-based models, yet the faithfulness and stability of explanation methods in non-English languages remain underexplored. This work presents a comprehensive benchmark of token-level XAI methods for Arabic sentiment analysis, evaluating LIME, SHAP, Integrated Gradients, DeepLIFT, and multiple ensemble variants across two transformer architectures (CAMeLBERT and AraBERT). We assess explanations using five established faithfulness metrics and complement score-based evaluation with rank-based aggregation via Borda count. We show that selective ensembling - particularly combining LIME and SHAP - yields a statistically significant but modest improvement over individual methods, improving ranking stability and robustness rather than absolute explanation quality. Bootstrap confidence intervals and paired Wilcoxon tests confirm the consistency of this effect. Our analysis further highlights persistent limitations in faithfulness metrics, including low correlation with Leave-One-Out perturbations, underscoring ongoing challenges in XAI evaluation. Overall, this study provides a rigorous, reproducible benchmark and practical guidance for explanation method selection in Arabic NLP. Explainable AI (XAI) Natural language processing Transformers Arabic sentiment analysis SHAP LIME Integrated Gradients DeepLIFT Full Text Additional Declarations No competing interests reported. Supplementary Files arabicsaxaibenchmark.zip Cite Share Download PDF Status: Posted Version 1 posted 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-8627067","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":576896566,"identity":"eeb5aa1b-3203-4ffd-a62f-a5295cb8d9ff","order_by":0,"name":"Youssef Chafiqui","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA5UlEQVRIie3PvQrCMBDA8ZNAuhy6Rvx8hECgCH0Zi6BLFEcHEUHopq/gKzg5R4p1qXv3goNTQRAHEZsiuLUdBfMf7gjkBwmAyfSTIcCMA1BrpdITa5YjoSYY9DXBkkRvJvnnWBA/nQ+JmjruBuTtGs17CJZ/3OWScDJgio9cD857Rwbpw3A4jHJJhLyRcN/1Kuu9kDQlDO0iIh5KE4IXIV/liM0yQpHEY68EqYfS7qV/ER5Sm4w3DGnRX6qnUETq6bS22zi+yfuiXbP8IJd0VbaIHpRlM++6rrP8EpIU3TaZTKb/7A1w70kozjKVXAAAAABJRU5ErkJggg==","orcid":"","institution":"Hassan II University","correspondingAuthor":true,"prefix":"","firstName":"Youssef","middleName":"","lastName":"Chafiqui","suffix":""},{"id":576896567,"identity":"4d0e025f-5232-45ed-bd2a-c74d933d48de","order_by":1,"name":"Houda Anoun","email":"","orcid":"","institution":"Hassan II University","correspondingAuthor":false,"prefix":"","firstName":"Houda","middleName":"","lastName":"Anoun","suffix":""}],"badges":[],"createdAt":"2026-01-17 15:38:40","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8627067/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8627067/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":100976725,"identity":"924e2455-a701-40f0-99f5-136008e82d3e","added_by":"auto","created_at":"2026-01-23 11:12:22","extension":"json","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":4219,"visible":true,"origin":"","legend":"","description":"","filename":"29340f7ef01344ce966927f05aa8ca9b.json","url":"https://assets-eu.researchsquare.com/files/rs-8627067/v1/803b26e93662ce05d9466ac7.json"},{"id":100976726,"identity":"9d6264f7-dd87-4e3f-84d6-85eeaa449002","added_by":"auto","created_at":"2026-01-23 11:12:22","extension":"xml","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":105625,"visible":true,"origin":"","legend":"","description":"","filename":"29340f7ef01344ce966927f05aa8ca9b1enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-8627067/v1/ed7df9ad51c74c1483ce4105.xml"},{"id":101297222,"identity":"7ce08035-cdbc-4b15-846b-2bf985c79a24","added_by":"auto","created_at":"2026-01-28 09:26:02","extension":"pdf","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":490058,"visible":true,"origin":"","legend":"","description":"","filename":"V2EvaluationBenchmarkStudyforXAIMethodsinArabicSentimentAnalysis.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8627067/v1/3dbc660654f924a197db689b.pdf"},{"id":100976727,"identity":"2f4d4e2c-1db4-4b8c-9c4c-9eae24b565fc","added_by":"auto","created_at":"2026-01-23 11:12:22","extension":"png","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":28769,"visible":true,"origin":"","legend":"","description":"","filename":"XAIEvaluationdiagram.png","url":"https://assets-eu.researchsquare.com/files/rs-8627067/v1/4c49759d43d8a1c55e2ea0b3.png"},{"id":100976747,"identity":"7fd8882b-4add-4d87-8e59-9502cfa51150","added_by":"auto","created_at":"2026-01-23 11:12:27","extension":"zip","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":103746612,"visible":true,"origin":"","legend":"","description":"","filename":"arabicsaxaibenchmark.zip","url":"https://assets-eu.researchsquare.com/files/rs-8627067/v1/1bb564f112d0a2dd5e5b0244.zip"},{"id":101397627,"identity":"0269e1cb-0071-41fd-922a-7a9ee5fb3925","added_by":"auto","created_at":"2026-01-29 09:33:15","extension":"png","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":46418,"visible":true,"origin":"","legend":"","description":"","filename":"elbowexample.png","url":"https://assets-eu.researchsquare.com/files/rs-8627067/v1/8c738d09ab5662380eb2be30.png"},{"id":100976728,"identity":"8724ed48-f1e2-49d6-a5ba-1eef2bd49276","added_by":"auto","created_at":"2026-01-23 11:12:22","extension":"png","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":68823,"visible":true,"origin":"","legend":"","description":"","filename":"lmethodexample.png","url":"https://assets-eu.researchsquare.com/files/rs-8627067/v1/b9dec7bdfc426bcefedb3179.png"},{"id":100976729,"identity":"96c7bf30-0021-418b-8ee7-44313743641b","added_by":"auto","created_at":"2026-01-23 11:12:22","extension":"bst","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":146013,"visible":true,"origin":"","legend":"","description":"","filename":"snapacite.bst","url":"https://assets-eu.researchsquare.com/files/rs-8627067/v1/e901d6eeed4a09a96fe47d1c.bst"},{"id":101297010,"identity":"a7820331-2f06-4f7b-a58d-e08390a4ebfc","added_by":"auto","created_at":"2026-01-28 09:24:25","extension":"bst","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":29828,"visible":true,"origin":"","legend":"","description":"","filename":"snaps.bst","url":"https://assets-eu.researchsquare.com/files/rs-8627067/v1/1bb63fd7c8200574c49b2643.bst"},{"id":100976731,"identity":"4363d4b4-3c70-4f37-9fa9-4da5daa6fe2c","added_by":"auto","created_at":"2026-01-23 11:12:22","extension":"bst","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":35515,"visible":true,"origin":"","legend":"","description":"","filename":"snbasic.bst","url":"https://assets-eu.researchsquare.com/files/rs-8627067/v1/ef4a497c2ae371ee6e60d46c.bst"},{"id":100976732,"identity":"2d1497da-1c78-48a2-8dc7-eee44118a937","added_by":"auto","created_at":"2026-01-23 11:12:22","extension":"bst","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":33968,"visible":true,"origin":"","legend":"","description":"","filename":"snchicago.bst","url":"https://assets-eu.researchsquare.com/files/rs-8627067/v1/435940a02ba6967f0eda7df4.bst"},{"id":101880547,"identity":"2cbea59c-dc26-43e5-961d-7c84ef68afc4","added_by":"auto","created_at":"2026-02-04 15:03:30","extension":"cls","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":55857,"visible":true,"origin":"","legend":"","description":"","filename":"snjnl.cls","url":"https://assets-eu.researchsquare.com/files/rs-8627067/v1/f2d4059d198cbf0eddbcf63c.cls"},{"id":100976740,"identity":"91eddd42-14aa-4807-96db-7446095c1b99","added_by":"auto","created_at":"2026-01-23 11:12:25","extension":"bst","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":64023,"visible":true,"origin":"","legend":"","description":"","filename":"snmathphysay.bst","url":"https://assets-eu.researchsquare.com/files/rs-8627067/v1/b82f2d10167c0535a84a604e.bst"},{"id":101751158,"identity":"235d8e8a-07c3-4960-a076-26edec08c579","added_by":"auto","created_at":"2026-02-03 10:17:01","extension":"bst","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":64166,"visible":true,"origin":"","legend":"","description":"","filename":"snmathphysnum.bst","url":"https://assets-eu.researchsquare.com/files/rs-8627067/v1/ec858655f81fe8e2756a0662.bst"},{"id":101203006,"identity":"cc415476-3ac2-4510-bbe3-dd1d354f90f4","added_by":"auto","created_at":"2026-01-27 09:38:30","extension":"bst","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":37333,"visible":true,"origin":"","legend":"","description":"","filename":"snnature.bst","url":"https://assets-eu.researchsquare.com/files/rs-8627067/v1/139d745304caa4039fb05b50.bst"},{"id":100976733,"identity":"96e5b3c8-e575-4050-8e5a-19266c513ce5","added_by":"auto","created_at":"2026-01-23 11:12:22","extension":"bst","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":39951,"visible":true,"origin":"","legend":"","description":"","filename":"snvancouveray.bst","url":"https://assets-eu.researchsquare.com/files/rs-8627067/v1/966d473fa8970d573ddd652a.bst"},{"id":100976742,"identity":"4bcdf2d4-94ca-4797-ae82-03f33c892d88","added_by":"auto","created_at":"2026-01-23 11:12:25","extension":"bst","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":40758,"visible":true,"origin":"","legend":"","description":"","filename":"snvancouvernum.bst","url":"https://assets-eu.researchsquare.com/files/rs-8627067/v1/b77e62087ff6e24b71e55c35.bst"},{"id":101203008,"identity":"80cebd79-df5f-4b27-b547-ce2dca2f6a87","added_by":"auto","created_at":"2026-01-27 09:38:30","extension":"xml","order_by":20,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":116675,"visible":true,"origin":"","legend":"","description":"","filename":"29340f7ef01344ce966927f05aa8ca9b1structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-8627067/v1/75477b72848c860d9ade2257.xml"},{"id":101203239,"identity":"9424cf27-14d2-41fa-9e00-c2d35ae94c18","added_by":"auto","created_at":"2026-01-27 09:39:09","extension":"html","order_by":21,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":130756,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-8627067/v1/138c08a5f1132baa42b4a6a7.html"},{"id":101882314,"identity":"7f779871-0620-4e69-8d5f-b79127c2eb0d","added_by":"auto","created_at":"2026-02-04 15:21:58","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":450944,"visible":true,"origin":"","legend":"","description":"","filename":"V2EvaluationBenchmarkStudyforXAIMethodsinArabicSentimentAnalysis.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8627067/v1_covered_1463d7b3-8ff8-49a8-b847-910525aa27c0.pdf"},{"id":100976745,"identity":"cd19adf1-f6da-4592-8d0f-e1846c4a1418","added_by":"auto","created_at":"2026-01-23 11:12:26","extension":"zip","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":103746612,"visible":true,"origin":"","legend":"","description":"","filename":"arabicsaxaibenchmark.zip","url":"https://assets-eu.researchsquare.com/files/rs-8627067/v1/a270780e2bc0f36c467d764b.zip"}],"financialInterests":"No competing interests reported.","formattedTitle":"Evaluation Benchmark Study for XAI Methods in Arabic Sentiment Analysis","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"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":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Explainable AI (XAI), Natural language processing, Transformers, Arabic sentiment analysis, SHAP, LIME, Integrated Gradients, DeepLIFT","lastPublishedDoi":"10.21203/rs.3.rs-8627067/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8627067/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eExplainable Artificial Intelligence (XAI) is essential for interpreting transformer-based models, yet the faithfulness and stability of explanation methods in non-English languages remain underexplored. This work presents a comprehensive benchmark of token-level XAI methods for Arabic sentiment analysis, evaluating LIME, SHAP, Integrated Gradients, DeepLIFT, and multiple ensemble variants across two transformer architectures (CAMeLBERT and AraBERT). We assess explanations using five established faithfulness metrics and complement score-based evaluation with rank-based aggregation via Borda count. We show that selective ensembling - particularly combining LIME and SHAP - yields a statistically significant but modest improvement over individual methods, improving ranking stability and robustness rather than absolute explanation quality. Bootstrap confidence intervals and paired Wilcoxon tests confirm the consistency of this effect. Our analysis further highlights persistent limitations in faithfulness metrics, including low correlation with Leave-One-Out perturbations, underscoring ongoing challenges in XAI evaluation. Overall, this study provides a rigorous, reproducible benchmark and practical guidance for explanation method selection in Arabic NLP.\u003c/p\u003e","manuscriptTitle":"Evaluation Benchmark Study for XAI Methods in Arabic Sentiment Analysis","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-01-23 11:12:05","doi":"10.21203/rs.3.rs-8627067/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"50a778c8-c378-4aa4-9c87-cc6b0cd5e148","owner":[],"postedDate":"January 23rd, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2026-01-23T11:12:05+00:00","versionOfRecord":[],"versionCreatedAt":"2026-01-23 11:12:05","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8627067","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8627067","identity":"rs-8627067","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}