Testing AI and ML tools for formalising unstructured medical texts.
Abstract
Project objective: Develop a solution for automatic processing of medical texts, identification and structuring of clinical terms using SNOMED CT codes and FHIR standard. Work plan: 1. Clinical term identification - Preprocessing of Estonian medical texts - Application of machine learning/language models for term identification from text - Structuring of identified terms for further processing 2. Standardization and coding - Linking identified terms with SNOMED CT codes - Semantic matching of data with standardized ontologies - Converting results to FHIR format Demo solution and expected outputs: - Practical demonstration solution for complete workflow - Working system for automatic identification of clinical terms from Estonian texts - Structured output of terms linked with SNOMED CT codes - Standard data presentation in FHIR format for further use Keywords: medical texts, SNOMED CT, FHIR, language processing, clinical terms, automatic annotation, structured data
Related Papers
Safety and quality of high-risk plant-based foods and meat alternatives
Roasto, Mati
The Circular Schools – Empowering Secondary Education Students for a Green Future through Circularity Thinking Strategies
Voronova, Viktoria
Developing Estonian startup ecosystem and startup incubation programs: Part 1 - Developing the deep-tech startup ecosystem.
Lööve, Triinu