Skip to Main content Skip to Navigation
Theses

Knowledge graphs based extension of patients’ files to predict hospitalization

Raphaël Gazzotti 1
1 WIMMICS - Web-Instrumented Man-Machine Interactions, Communities and Semantics
CRISAM - Inria Sophia Antipolis - Méditerranée , Laboratoire I3S - SPARKS - Scalable and Pervasive softwARe and Knowledge Systems
Abstract : The use of electronic medical records (EMRs) and electronic prescribing are priorities in the various European action plans on connected health. The development of the EMR is a tremendous source of data; it captures all symptomatic episodes in a patient’s life and should lead to improved medical and care practices, as long as automatic treatment procedures are set up. As such, we are working on hospitalization prediction based on EMRs and after having represented them in vector form, we enrich these models in order to benefit from the knowledge resulting from referentials, whether generalist or specific in the medical field, in order to improve the predictive power of automatic classification algorithms. Determining the knowledge to be extracted with the objective of integrating it into vector representations is both a subjective task and intended for experts, we will see a semi-supervised procedure to partially automate this process. As a result of our research, we designed a product for general practitioners to prevent their patients from being hospitalized or at least improve their health. Thus, through a simulation, it will be possible for the doctor to evaluate the factors involved on the risk of hospitalization of his patient and to define the preventive actions to be planned to avoid the occurrence of this event. This decision support algorithm is intended to be directly integrated into the physician consultation software. For this purpose, we have developed in collaboration with many professional bodies, including the first to be concerned, general practitioners.
Complete list of metadata

https://tel.archives-ouvertes.fr/tel-03135236
Contributor : Abes Star :  Contact
Submitted on : Monday, February 8, 2021 - 6:56:08 PM
Last modification on : Friday, May 7, 2021 - 11:17:01 AM

File

2020COAZ4018.pdf
Version validated by the jury (STAR)

Identifiers

  • HAL Id : tel-03135236, version 3

Citation

Raphaël Gazzotti. Knowledge graphs based extension of patients’ files to predict hospitalization. Modeling and Simulation. Université Côte d'Azur, 2020. English. ⟨NNT : 2020COAZ4018⟩. ⟨tel-03135236v3⟩

Share

Metrics

Record views

182

Files downloads

299