This text discusses a project for a concept design that aims to help doctors better diagnose and treat Gastroesophageal Reflux Disease (GERD) using the latest methods in artificial intelligence (AI), specifically LangChain and ChatGPT-4 from OpenAI. The project includes creating a system for GERD, using AI to work with medical data, testing the system with some patient cases, and seeing how well it helps in treating GERD. The project seeks to explore the feasibility of implementing a system that enhances the diagnosis and treatment through the application. 50 cases were generated for evaluation, with a ChatGPT response for each case. Of these, 36 cases were accurately evaluated with correct references, while 14 cases received satisfactory evaluations but included references not originally provided to ChatGPT. Overall, the recommendations made by ChatGPT were deemed satisfactory, showcasing the model's capability to offer preliminary advice and insights
Eser Adı (dc.title) | Natural Language Processing and Machine Learning-Based Medical Decision Support Application Concept Design for Gastroesophageal Reflux Disease |
Eser Sahibi (dc.contributor.author) | Tevfik Aşgın |
Tez Danışmanı (dc.contributor.advisor) | Sıla Övgü Korkut Uysal |
Yayıncı (dc.publisher) | İzmir Katip Çelebi Üniversitesi Fen Bilimleri Enstitüsü |
Tür (dc.type) | Diğer |
Özet (dc.description.abstract) | This text discusses a project for a concept design that aims to help doctors better diagnose and treat Gastroesophageal Reflux Disease (GERD) using the latest methods in artificial intelligence (AI), specifically LangChain and ChatGPT-4 from OpenAI. The project includes creating a system for GERD, using AI to work with medical data, testing the system with some patient cases, and seeing how well it helps in treating GERD. The project seeks to explore the feasibility of implementing a system that enhances the diagnosis and treatment through the application. 50 cases were generated for evaluation, with a ChatGPT response for each case. Of these, 36 cases were accurately evaluated with correct references, while 14 cases received satisfactory evaluations but included references not originally provided to ChatGPT. Overall, the recommendations made by ChatGPT were deemed satisfactory, showcasing the model's capability to offer preliminary advice and insights |
Kayıt Giriş Tarihi (dc.date.accessioned) | 2024-03-13 |
Açık Erişim Tarihi (dc.date.available) | 2024-03-13 |
Yayın Tarihi (dc.date.issued) | 2024 |
Tek Biçim Adres (dc.identifier.uri) | https://hdl.handle.net/11469/3955 |
Yayın Dili (dc.language.iso) | eng |
Konu Başlıkları (dc.subject) | LangChain |
Konu Başlıkları (dc.subject) | ChatGPT-4 |
Haklar (dc.rights) | Open access |