{"id":1826,"date":"2026-03-24T00:06:22","date_gmt":"2026-03-24T00:06:22","guid":{"rendered":"https:\/\/cloudlab.urv.cat\/catedracloud\/?p=1826"},"modified":"2026-03-24T11:26:12","modified_gmt":"2026-03-24T11:26:12","slug":"hackato-ia-urv-tsystems-2026-sia","status":"publish","type":"post","link":"https:\/\/cloudlab.urv.cat\/catedracloud\/2026\/03\/24\/hackato-ia-urv-tsystems-2026-sia\/","title":{"rendered":"Hackat\u00f3 IA URV-TSystems 2026: SIA"},"content":{"rendered":"\n<p><\/p>\n\n\n\n<div class=\"wp-block-uagb-image uagb-block-acbf9101 wp-block-uagb-image--layout-default wp-block-uagb-image--effect-static wp-block-uagb-image--align-none\"><figure class=\"wp-block-uagb-image__figure\"><img decoding=\"async\" srcset=\"https:\/\/cloudlab.urv.cat\/catedracloud\/wp-content\/uploads\/2026\/03\/D81_3478-1024x684.jpg ,https:\/\/cloudlab.urv.cat\/catedracloud\/wp-content\/uploads\/2026\/03\/D81_3478.jpg 780w, https:\/\/cloudlab.urv.cat\/catedracloud\/wp-content\/uploads\/2026\/03\/D81_3478.jpg 360w\" sizes=\"auto, (max-width: 480px) 150px\" src=\"https:\/\/cloudlab.urv.cat\/catedracloud\/wp-content\/uploads\/2026\/03\/D81_3478-1024x684.jpg\" alt=\"\" class=\"uag-image-1827\" width=\"2048\" height=\"1367\" title=\"D81_3478\" loading=\"lazy\" role=\"img\"\/><\/figure><\/div>\n\n\n\n<p>SIA \u00e9s un sistema d&#8217;IA que classifica autom\u00e0ticament pacients com a NO cr\u00f2nics, PCC o MACA a partir de dades cl\u00edniques reals. Combina models d&#8217;aprenentatge autom\u00e0tic, cerca de similitud amb FAISS i un LLM local per generar informes cl\u00ednics i consells personalitzats.<\/p>\n\n\n\n<p><strong>Nom Grup:<\/strong> TarrIAco<\/p>\n\n\n\n<p>Entitat del repte: C\u00e0ritas Diocesana de Tarragona\/C\u00e0tedra UNESCO d&#8217;Habitatge de la URV\/Hospital Universitari Joan XXIII<\/p>\n\n\n\n<p><strong>Integrants:<\/strong><\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>Gen\u00eds Aragon\u00e8s Torralbo (<a>genisarato@gmail.com<\/a>)<\/li>\n\n\n\n<li>Massin Laaouaj (<a>massin.laaouaj@estudiants.urv.cat<\/a>)<\/li>\n\n\n\n<li>David Quintana Palomar (<a>dqp220504@gmail.com<\/a>)<\/li>\n\n\n\n<li>Marc Roda Cort\u00e9s (<a>1516marcroda@gmail.com<\/a>)<\/li>\n<\/ul>\n\n\n\n<p><strong>Nom Projecte:<\/strong> SIA (Salut en Intelig\u00e8ncia Artificial)<\/p>\n\n\n\n<p><strong>Descripci\u00f3:<\/strong> SIA neix de la necessitat de donar suport als professionals sanitaris de l\u2019Hospital Joan XXIII en la complexa tasca d\u2019identificar i categoritzar pacients cr\u00f2nics. El sistema automatitza la classificaci\u00f3 triclasse (NO \/ PCC \/ MACA) a partir de m\u00e9s de 80 variables cl\u00edniques per pacient, incloent dades demogr\u00e0fiques, diagn\u00f2stics, f\u00e0rmacs prescrits, visites a urg\u00e8ncies i hospitalitzacions. La soluci\u00f3 es basa en dues intel\u00b7lig\u00e8ncies artificials: d\u2019una banda, un nucli predictiu amb un pipeline jer\u00e0rquic en dos estadis, on un primer model (Random Forest) determina si un pacient \u00e9s cr\u00f2nic o no, i un segon model (HistGradientBoosting calibrat) diferencia entre PCC i MACA, millorant la precisi\u00f3 respecte a un classificador directe; de l\u2019altra, un sistema de cerca de similitud basat en FAISS que permet identificar en mil\u00b7lisegons els pacients m\u00e9s semblants, proporcionant evid\u00e8ncia estad\u00edstica \u00fatil per a la presa de decisions cl\u00edniques. A m\u00e9s, un model de llenguatge local (Llama 3.1:8b via Ollama) genera tant informes cl\u00ednics t\u00e8cnics per als professionals com recomanacions personalitzades per als pacients en catal\u00e0. Tot el sistema s\u2019exposa mitjan\u00e7ant una API REST amb Flask i es consumeix des d\u2019un portal web desenvolupat amb Laravel, separant clarament el backend d\u2019IA del frontend sanitari.<br><\/p>\n\n\n\n<p><strong>Reptes abordats:<\/strong><\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>Identificaci\u00f3 autom\u00e0tica de pacients en categories de cronicitat: **NO cr\u00f2nic**, **PCC** (Pacient Cr\u00f2nic Complex) i **MACA** (Malaltia Avan\u00e7ada en Congesti\u00f3 Alta).<\/li>\n\n\n\n<li>Cerca de pacients similars per donar suport als cl\u00ednics en la presa de decisions basada en evid\u00e8ncia.<\/li>\n\n\n\n<li>Generaci\u00f3 d&#8217;informes cl\u00ednics automatitzats amb IA generativa (LLM local).<\/li>\n\n\n\n<li>Portal web per facilitar l&#8217;acc\u00e9s dels professionals sanitaris i dels propis pacients a la informaci\u00f3 personalitzada.<\/li>\n<\/ul>\n\n\n\n<p><strong>Tecnologies utilitzades:<\/strong><\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>Llenguatges de programaci\u00f3<\/strong>: \n<ul class=\"wp-block-list\">\n<li>Python 3.11+ <\/li>\n\n\n\n<li>PHP 8.2 (Laravel) <\/li>\n\n\n\n<li>JavaScript (ES6+)<\/li>\n\n\n\n<li>Blade (Laravel templates) <\/li>\n<\/ul>\n<\/li>\n\n\n\n<li><strong>Frameworks i llibreries<\/strong>: \n<ul class=\"wp-block-list\">\n<li><strong>**Scikit-learn**<\/strong>: RandomForestClassifier, HistGradientBoostingClassifier, CalibratedClassifierCV, pipeline de preprocessament <\/li>\n\n\n\n<li><strong>FAISS<\/strong> (faiss-cpu): \u00edndex de similitud vetorial per a cerca de ve\u00efns propers <\/li>\n\n\n\n<li><strong>Flask + Flask-CORS<\/strong>: API REST del backend d&#8217;IA <\/li>\n\n\n\n<li><strong>Ollama (Llama 3.1:8b)<\/strong>: LLM local per a generaci\u00f3 d&#8217;informes cl\u00ednics i consells al pacient <\/li>\n\n\n\n<li><strong>Pandas \/ NumPy<\/strong>: transformaci\u00f3 i processament del dataset <\/li>\n\n\n\n<li><strong>Joblib**<\/strong>: serialitzaci\u00f3 dels models entrenats &#8211; <\/li>\n\n\n\n<li><strong>Laravel 11**<\/strong>: portal web sanitari amb autenticaci\u00f3 i gesti\u00f3 de rols <\/li>\n<\/ul>\n<\/li>\n\n\n\n<li><strong>Eines i plataformes<\/strong> \n<ul class=\"wp-block-list\">\n<li><strong>Supabase (PostgreSQL)<\/strong>: base de dades cl\u00ednica de l&#8217;hospital (acc\u00e9s read-only) &#8211; <\/li>\n\n\n\n<li><strong>Git \/ GitHub<\/strong>: control de versions i treball col\u00b7laboratiu &#8211;<\/li>\n\n\n\n<li><strong>Python venv<\/strong>: entorn virtual per al backend de Python &#8211; <\/li>\n\n\n\n<li><strong>Composer \/ npm<\/strong>: gesti\u00f3 de depend\u00e8ncies del portal Laravel<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n\n\n\n<p><strong>Recursos del Projecte:<\/strong><\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>Github:<\/strong> <a href=\"https:\/\/github.com\/Genisarato\/TarrIAco\">https:\/\/github.com\/Genisarato\/TarrIAco<\/a><\/li>\n\n\n\n<li><strong>V\u00eddeo:<\/strong> <a href=\"https:\/\/rovira-my.sharepoint.com\/:v:\/g\/personal\/47863618-c_epp_urv_cat\/IQDq1poqA-t_S4pssBkdDXegAYCCqVmUk-bgWmiWwMXRAfs?e=bnWpQJ\">https:\/\/rovira-my.sharepoint.com\/:v:\/g\/personal\/47863618-c_epp_urv_cat\/IQDq1poqA-t_S4pssBkdDXegAYCCqVmUk-bgWmiWwMXRAfs?e=bnWpQJ<\/a><\/li>\n\n\n\n<li><strong>Presentaci\u00f3: <\/strong><a href=\"https:\/\/rovira-my.sharepoint.com\/:b:\/g\/personal\/52460503-x_epp_urv_cat\/IQAw9365eFE5R7I4MJ2IhG2FAfR3jFdK4kh5QMgMeRSIvE0?e=EmQ1X8\">https:\/\/rovira-my.sharepoint.com\/:b:\/g\/personal\/52460503-x_epp_urv_cat\/IQAw9365eFE5R7I4MJ2IhG2FAfR3jFdK4kh5QMgMeRSIvE0?e=EmQ1X8<\/a><\/li>\n<\/ul>\n\n\n\n<p><strong>Acc\u00e9s a la resta de projectes de la Hackat\u00f3 d&#8217;IA 2026:<\/strong> <a href=\"https:\/\/cloudlab.urv.cat\/catedracloud\/hackato2026\/\">https:\/\/cloudlab.urv.cat\/catedracloud\/hackato2026\/<\/a><\/p>\n\n\n\n<p><\/p>\n","protected":false},"excerpt":{"rendered":"<p>SIA \u00e9s un sistema d&#8217;IA que classifica autom\u00e0ticament pacients com a NO cr\u00f2nics, PCC o MACA a partir de dades cl\u00edniques reals. Combina models d&#8217;aprenentatge autom\u00e0tic, cerca de similitud amb FAISS i un LLM local per generar informes cl\u00ednics i consells personalitzats. Nom Grup: TarrIAco Entitat del repte: C\u00e0ritas Diocesana de Tarragona\/C\u00e0tedra UNESCO d&#8217;Habitatge de [&hellip;]<\/p>\n","protected":false},"author":2,"featured_media":1828,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_uag_custom_page_level_css":"","_swt_meta_header_display":false,"_swt_meta_footer_display":false,"_swt_meta_site_title_display":false,"_swt_meta_sticky_header":false,"_swt_meta_transparent_header":false,"footnotes":""},"categories":[113,94,39],"tags":[],"class_list":["post-1826","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-cloud-computing","category-hackathon","category-ia"],"jetpack_featured_media_url":"https:\/\/cloudlab.urv.cat\/catedracloud\/wp-content\/uploads\/2026\/03\/Captura-de-pantalla-2026-03-24-a-las-1.04.13.png","uagb_featured_image_src":{"full":["https:\/\/cloudlab.urv.cat\/catedracloud\/wp-content\/uploads\/2026\/03\/Captura-de-pantalla-2026-03-24-a-las-1.04.13.png",1178,724,false],"thumbnail":["https:\/\/cloudlab.urv.cat\/catedracloud\/wp-content\/uploads\/2026\/03\/Captura-de-pantalla-2026-03-24-a-las-1.04.13-150x150.png",150,150,true],"medium":["https:\/\/cloudlab.urv.cat\/catedracloud\/wp-content\/uploads\/2026\/03\/Captura-de-pantalla-2026-03-24-a-las-1.04.13-300x184.png",300,184,true],"medium_large":["https:\/\/cloudlab.urv.cat\/catedracloud\/wp-content\/uploads\/2026\/03\/Captura-de-pantalla-2026-03-24-a-las-1.04.13-768x472.png",768,472,true],"large":["https:\/\/cloudlab.urv.cat\/catedracloud\/wp-content\/uploads\/2026\/03\/Captura-de-pantalla-2026-03-24-a-las-1.04.13-1024x629.png",1024,629,true],"1536x1536":["https:\/\/cloudlab.urv.cat\/catedracloud\/wp-content\/uploads\/2026\/03\/Captura-de-pantalla-2026-03-24-a-las-1.04.13.png",1178,724,false],"2048x2048":["https:\/\/cloudlab.urv.cat\/catedracloud\/wp-content\/uploads\/2026\/03\/Captura-de-pantalla-2026-03-24-a-las-1.04.13.png",1178,724,false]},"uagb_author_info":{"display_name":"Carlos Molina","author_link":"https:\/\/cloudlab.urv.cat\/catedracloud\/author\/carlos-molina\/"},"uagb_comment_info":0,"uagb_excerpt":"SIA \u00e9s un sistema d&#8217;IA que classifica autom\u00e0ticament pacients com a NO cr\u00f2nics, PCC o MACA a partir de dades cl\u00edniques reals. Combina models d&#8217;aprenentatge autom\u00e0tic, cerca de similitud amb FAISS i un LLM local per generar informes cl\u00ednics i consells personalitzats. Nom Grup: TarrIAco Entitat del repte: C\u00e0ritas Diocesana de Tarragona\/C\u00e0tedra UNESCO d&#8217;Habitatge de&hellip;","_links":{"self":[{"href":"https:\/\/cloudlab.urv.cat\/catedracloud\/wp-json\/wp\/v2\/posts\/1826","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/cloudlab.urv.cat\/catedracloud\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/cloudlab.urv.cat\/catedracloud\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/cloudlab.urv.cat\/catedracloud\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/cloudlab.urv.cat\/catedracloud\/wp-json\/wp\/v2\/comments?post=1826"}],"version-history":[{"count":6,"href":"https:\/\/cloudlab.urv.cat\/catedracloud\/wp-json\/wp\/v2\/posts\/1826\/revisions"}],"predecessor-version":[{"id":1903,"href":"https:\/\/cloudlab.urv.cat\/catedracloud\/wp-json\/wp\/v2\/posts\/1826\/revisions\/1903"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/cloudlab.urv.cat\/catedracloud\/wp-json\/wp\/v2\/media\/1828"}],"wp:attachment":[{"href":"https:\/\/cloudlab.urv.cat\/catedracloud\/wp-json\/wp\/v2\/media?parent=1826"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/cloudlab.urv.cat\/catedracloud\/wp-json\/wp\/v2\/categories?post=1826"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/cloudlab.urv.cat\/catedracloud\/wp-json\/wp\/v2\/tags?post=1826"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}