Comparative analysis of artificial intelligence algorithms for fine particulate matter prediction

Comparative analysis of artificial intelligence algorithms for fine particulate matter prediction | 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 Article Comparative analysis of artificial intelligence algorithms for fine particulate matter prediction Martín Rodríguez Núñez, Mónica Balzarini This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7199213/v1 This work is licensed under a CC BY 4.0 License Status: Under Revision Version 1 posted 14 You are reading this latest preprint version Abstract This study aims to analyze the predictive performance of artificial intelligence algorithms in forecasting fine particulate matter (PM 2.5 ) under different scenarios of temporal variability. PM 2.5 data were collected over a year using low-cost sensors in ten sites, under urban and agricultural land use in central Argentina. Additionally, 178 time series of satellite variables were downloaded from the cloud to be used as predictors. Various machine learning models, including Linear Regression, Random Forest, and XGBoost, as well as deep learning models such as Long Short-Term Memory (LSTM), Gated Recurrent Units (GRUs) and Recurrent Neural Network (RNN), were compared. The hourly concentration of PM 2.5 ranged from 0.53 \(\:\frac{\mu\:g}{{m}^{3}}\) to 95.28 \(\:\frac{\mu\:g}{{m}^{3}}\) , with an average of 13.1 \(\:\frac{\mu\:g}{{m}^{3}}\) for the urban land use and 6.78 \(\:\frac{\mu\:g}{{m}^{3}}\) in the agricultural one. Temporal variability was highly different between scenarios, the interquantilic ranges were 7.31 and 4.38 \(\:\frac{\mu\:g}{{m}^{3}}\) for the urban and agricultural land use, respectively. The GRU was the best algorithm in the urban land use (RMSE = 3.23 \(\:\frac{\mu\:g}{{m}^{3}}\) ). Conversely, for agricultural land use, the RF algorithm exhibited superior performance (RMSE = 2.54 \(\:\frac{\mu\:g}{{m}^{3}}\) ). The study findings highlight the strong impact of the time series variability and the learning capabilities of each predictive algorithm on predictions. Deep learning algorithms outperformed machine learning models in predicting PM 2.5 from highly variable time series. Artificial intelligence algorithms demonstrate the capability for accurate forecasting of PM 2.5 concentrations in the context of a low-cost alert system. Earth and environmental sciences/Climate sciences Earth and environmental sciences/Environmental sciences Physical sciences/Mathematics and computing PM2.5 exposure satellite data machine learning deep learning Full Text Additional Declarations No competing interests reported. Supplementary Files SupplementaryMaterial.docx GraphicalAbstract.tiff Cite Share Download PDF Status: Under Revision Version 1 posted Editorial decision: Revision requested 03 Sep, 2025 Reviews received at journal 02 Sep, 2025 Reviews received at journal 20 Aug, 2025 Reviews received at journal 16 Aug, 2025 Reviewers agreed at journal 08 Aug, 2025 Reviewers agreed at journal 05 Aug, 2025 Reviewers agreed at journal 03 Aug, 2025 Reviewers agreed at journal 03 Aug, 2025 Reviewers agreed at journal 03 Aug, 2025 Reviewers invited by journal 03 Aug, 2025 Editor invited by journal 29 Jul, 2025 Editor assigned by journal 25 Jul, 2025 Submission checks completed at journal 24 Jul, 2025 First submitted to journal 23 Jul, 2025 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-7199213","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":495047545,"identity":"1953b305-9bee-414d-9b0b-91fd6f947483","order_by":0,"name":"Martín Rodríguez Núñez","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAyUlEQVRIiWNgGAWjYLCCBBDB3gAkDCyI0sDYANbCcwCkRYJILWBKAqyRCC267WefP3i4416ewc3nVzf8KJBg4G/vTsCrxexMumFD4pniYoPbOWU3e4AOkzhzdgN+LQfSGBsS2xISN9zOSbvBA9RiIJFLQMv5Z1AtN8+k3fxDlJYbMFtusB+7TZwtN54xzgBpmXkmh+22jIEED2G/nE9j+PgTqKXv+PFnN9/8sZHjb+/FrwUJ8BiASWKVgwD7A1JUj4JRMApGwQgCANQuTd41DahPAAAAAElFTkSuQmCC","orcid":"","institution":"National University of Córdoba","correspondingAuthor":true,"prefix":"","firstName":"Martín","middleName":"Rodríguez","lastName":"Núñez","suffix":""},{"id":495047547,"identity":"f8629384-55ac-49b4-9f05-dce8af445200","order_by":1,"name":"Mónica Balzarini","email":"","orcid":"","institution":"National University of Córdoba","correspondingAuthor":false,"prefix":"","firstName":"Mónica","middleName":"","lastName":"Balzarini","suffix":""}],"badges":[],"createdAt":"2025-07-23 18:53:14","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7199213/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7199213/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":88363812,"identity":"6721db78-2a23-4b38-898f-b3295e6eb769","added_by":"auto","created_at":"2025-08-05 16:58:12","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":466932,"visible":true,"origin":"","legend":"","description":"","filename":"Paper.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7199213/v1_covered_310bf804-6500-4459-a90f-fee75ae386ba.pdf"},{"id":88362098,"identity":"562c1fe8-3f86-40c7-bfec-7d16f31283ba","added_by":"auto","created_at":"2025-08-05 16:34:10","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":38931,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryMaterial.docx","url":"https://assets-eu.researchsquare.com/files/rs-7199213/v1/f00e2aaff729ecb636de0854.docx"},{"id":88362107,"identity":"8abaa6ff-05a4-4d46-a8c5-ab50a9d64e35","added_by":"auto","created_at":"2025-08-05 16:34:10","extension":"tiff","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":921444,"visible":true,"origin":"","legend":"","description":"","filename":"GraphicalAbstract.tiff","url":"https://assets-eu.researchsquare.com/files/rs-7199213/v1/4e8d7b362e159d7a939ef853.tiff"}],"financialInterests":"No competing interests reported.","formattedTitle":"Comparative analysis of artificial intelligence algorithms for fine particulate matter prediction","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":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"PM2.5 exposure, satellite data, machine learning, deep learning","lastPublishedDoi":"10.21203/rs.3.rs-7199213/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7199213/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eThis study aims to analyze the predictive performance of artificial intelligence algorithms in forecasting fine particulate matter (PM\u003csub\u003e2.5\u003c/sub\u003e) under different scenarios of temporal variability. PM\u003csub\u003e2.5\u003c/sub\u003e data were collected over a year using low-cost sensors in ten sites, under urban and agricultural land use in central Argentina. Additionally, 178 time series of satellite variables were downloaded from the cloud to be used as predictors. Various machine learning models, including Linear Regression, Random Forest, and XGBoost, as well as deep learning models such as Long Short-Term Memory (LSTM), Gated Recurrent Units (GRUs) and Recurrent Neural Network (RNN), were compared. The hourly concentration of PM\u003csub\u003e2.5\u003c/sub\u003e ranged from 0.53 \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\frac{\\mu\\:g}{{m}^{3}}\\)\u003c/span\u003e\u003c/span\u003e to 95.28 \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\frac{\\mu\\:g}{{m}^{3}}\\)\u003c/span\u003e\u003c/span\u003e, with an average of 13.1 \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\frac{\\mu\\:g}{{m}^{3}}\\)\u003c/span\u003e\u003c/span\u003e for the urban land use and 6.78 \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\frac{\\mu\\:g}{{m}^{3}}\\)\u003c/span\u003e\u003c/span\u003e in the agricultural one. Temporal variability was highly different between scenarios, the interquantilic ranges were 7.31 and 4.38 \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\frac{\\mu\\:g}{{m}^{3}}\\)\u003c/span\u003e\u003c/span\u003e for the urban and agricultural land use, respectively. The GRU was the best algorithm in the urban land use (RMSE\u0026thinsp;=\u0026thinsp;3.23 \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\frac{\\mu\\:g}{{m}^{3}}\\)\u003c/span\u003e\u003c/span\u003e). Conversely, for agricultural land use, the RF algorithm exhibited superior performance (RMSE\u0026thinsp;=\u0026thinsp;2.54 \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\frac{\\mu\\:g}{{m}^{3}}\\)\u003c/span\u003e\u003c/span\u003e). The study findings highlight the strong impact of the time series variability and the learning capabilities of each predictive algorithm on predictions. Deep learning algorithms outperformed machine learning models in predicting PM\u003csub\u003e2.5\u003c/sub\u003e from highly variable time series. Artificial intelligence algorithms demonstrate the capability for accurate forecasting of PM\u003csub\u003e2.5\u003c/sub\u003e concentrations in the context of a low-cost alert system.\u003c/p\u003e","manuscriptTitle":"Comparative analysis of artificial intelligence algorithms for fine particulate matter prediction","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-08-05 16:34:06","doi":"10.21203/rs.3.rs-7199213/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2025-09-03T16:40:07+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-09-02T10:42:24+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-08-20T10:17:14+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-08-16T06:01:58+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"239785966617758529574181016250264678207","date":"2025-08-08T13:01:39+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"274995826165376785794320192094474198535","date":"2025-08-05T12:45:58+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"261907447568862213255319278687079766892","date":"2025-08-04T01:48:33+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"102059721056481042646470601487635418998","date":"2025-08-03T14:50:34+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"236789471226985723889737386990870024018","date":"2025-08-03T12:40:36+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-08-03T12:28:03+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2025-07-29T04:43:38+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-07-25T10:54:59+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-07-24T06:21:01+00:00","index":"","fulltext":""},{"type":"submitted","content":"Scientific Reports","date":"2025-07-23T18:42:07+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"7715139b-8e48-4b9c-b048-7465aefd0f3c","owner":[],"postedDate":"August 5th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"in-revision","subjectAreas":[{"id":52568832,"name":"Earth and environmental sciences/Climate sciences"},{"id":52568833,"name":"Earth and environmental sciences/Environmental sciences"},{"id":52568834,"name":"Physical sciences/Mathematics and computing"}],"tags":[],"updatedAt":"2025-09-03T16:53:15+00:00","versionOfRecord":[],"versionCreatedAt":"2025-08-05 16:34:06","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7199213","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7199213","identity":"rs-7199213","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}