FWO-ARRS projekt

(2019-2022)

Apr 1, 2020
Iz preprinta naprednega preglednega člnaka

Izboljšanje razumljivosti in napovedne zmogljivosti modelov za oceno tveganja in podporo odločanju v kliničnem okolju

Glavna institucija gostiteljica: KU Leuven
Vodja: Katrien Verbert (KU Leuven, Odsek za informatiko)

Tuja institucija gostiteljica: Univerza v Mariboru
Vodja: Gregor Stiglic (UM, Fakulteta za zdravstvene vede)

Spletna referenca na Informacijskem sistemu o raziskovalni dejavnosti v Sloveniji (SICRIS).

Povzetek
Nedavno ponovno odkrite in močno izboljšane metode nevronskih mrež, imenovane globoke nevronske mreže, postajajo vse bolj priljubljene kot napovedno orodje na različnih področjih zdravstva. Pokazalo se je, da globoke nevronske mreže večkrat presegajo manj zapletene modele napovedovanja, zlasti v primerih kjer je potreben inženiring spremenljivk. Po drugi strani pa ocena tveganja na podlagi globokih nevronskih mrež omogoča zelo omejeno interpretacijo dobljenih rezultatov s strani zdravstvenih strokovnjakov ali pacientov. Za reševanje tega izziva bo projekt raziskal vložitev modelov globokih nevronskih mrež in izvzem spremenljivk z metodami vizualne analitike. Vizualna analitika (VA) združuje avtomatizirano analizo z metodami vizualizacije, da bi dobili vpogled v zapletena podatkovja in interpretabilnost modelov. Natančneje bomo raziskali, kako lahko oblikujemo, razvijamo in ocenjujemo interaktivne tehnike vizualne analitike na podlagi različnih napovednih modelov za oceno tveganja pri kroničnih bolezni, kot sta diabetes tipa 2 (DT2) in kardiovaskularne bolezni (KVD).

Splošni cilj je dvojen:

  1. pojasniti napovedi in negotovost teh napovedi zdravstvenim strokovnjakom, ki bo služila kot podpora pri odločanju in povečanje zaupanja uporabnikov ter

  2. vključitev vnosa in povratnih informacij končnih uporabnikov v postopek predvidevanja, kot so manjkajoče laboratorijske vrednosti in optimizacija parametrov. Pristop bo ovrednoten v dveh podrobnih študijah uporabnikov v kliničnih okoljih.

Interpretabilnost Napovedni modeli Vizualna analitika
Aditya Bhattacharya
Aditya Bhattacharya
PhD Student

Sorodno

Publikacije

Directive Explanations for Monitoring the Risk of Diabetes Onset: Introducing Directive Data-Centric Explanations and Combinations to Support What-If Explorations

Explainable artificial intelligence is increasingly used in machine learning (ML) based decision-making systems in healthcare. However, …
Koda Projekt DOI

Generating Extremely Short Summaries from the Scientific Literature to Support Decisions in Primary Healthcare: A Human Evaluation Study

Recent advancements in Natural Language Processing (NLP) using large pre-trained neural language models were recently used in various …
Koda Projekt DOI

The Impact of Mobile Health Use on the Self-care of Patients With Type 2 Diabetes: Protocol for a Randomized Controlled Trial

Background: Chronic diseases are a substantial public health issue worldwide and affect an individual’s quality of life. Due to the …
PDF Projekt DOI

Relevance of automated generated short summaries of scientific abstract: use case scenario in healthcare

The recent development and successful deployment of large pre-trained natural language models in few-shot and zero-shot scenarios …
PDF Koda Projekt DOI

Extracting New Temporal Features to Improve the Interpretability of Undiagnosed Type 2 Diabetes Mellitus Prediction Models

Type 2 diabetes mellitus (T2DM) often results in high morbidity and mortality. In addition, T2DM presents a substantial financial …
PDF Projekt DOI

Internet Use and Psychosomatic Symptoms among University Students: Cross-Sectional Study

Background: Although the internet facilitates access to a wide range of knowledge and evidence, overuse among young people is …
PDF Projekt DOI

Digital Tools in Behavior Change Support Education in Health and Other Students: A Systematic Review

Due to the increased prevalence of chronic diseases, behavior changes are integral to self-management. Healthcare and other …
PDF Projekt DOI

A Review of Mortality Risk Prediction Models in Smartphone Applications

Healthcare professionals in healthcare systems need access to freely available, real-time, evidence-based mortality risk prediction …
PDF Projekt DOI

Explaining artificial intelligence with visual analytics in healthcare

To make predictions and explore large datasets, healthcare is increasingly applying advanced algorithms of artificial intelligence. …
PDF Projekt DOI

Perception of the Online Learning Environment of Nursing Students in Slovenia: Validation of the DREEM Questionnaire

At the time of the outbreak of the coronavirus pandemic, several measures were in place to limit the spread of the virus, such as …
PDF Projekt DOI

Evaluating Quality, Usability, Evidence-Based Content, and Gamification Features in Mobile Learning Apps Designed to Teach Children Basic Life Support: Systematic Search in App Stores and Content Analysis

Background: Globally, 3.7 million people die of sudden cardiac death annually. Following the World Health Organization endorsement of …
PDF Projekt DOI

Health Recommender Systems: Systematic Review

Background: Health recommender systems (HRSs) offer the potential to motivate and engage users to change their behavior by sharing …
PDF Projekt DOI

Development and validation of the type 2 diabetes mellitus 10-year risk score prediction models from survey data

In this paper, we demonstrate the development and validation of the 10-years type 2 diabetes mellitus (T2DM) risk prediction models …
PDF Koda Projekt DOI

Emotional Intelligence among Nursing Students: Findings from a Longitudinal Study

Emotional intelligence is an important factor for nursing students’ success and work performance. Although the level of emotional …
PDF Projekt DOI

Local Interpretability of Calibrated Prediction Models: A Case of Type 2 Diabetes Mellitus Screening Test

Machine Learning (ML) models are often complex and difficult to interpret due to their “black-box” characteristics. Interpretability of …
PDF Koda Projekt Poster DOI

Early detection of type 2 diabetes mellitus using machine learning-based prediction models

There is a need of ensuring machine learning models that are interpretable. Higher interpretability of the model means easier …
PDF Koda Projekt

Interpreting influence of feature ranking in derivation of prediction models for screening questionnaires optimization

Questionnaire based screening tests have been widely used in different fields ranging from healthcare and psychology to business …
PDF Koda Projekt

The Effects of Gamification and Oral Self-care on Oral Hygiene in Children: a Systematic Review of mHealth Apps

Background: Poor oral hygiene care is a significant public health problem worldwide. Oral healthcare education is a public health …
PDF Projekt DOI

Interpretability of machine learning based prediction models in healthcare

There is a need of ensuring machine learning models that are interpretable. Higher interpretability of the model means easier …
PDF Projekt DOI

Local vs. global interpretability of machine learning models in type 2 diabetes mellitus screening

Machine learning based predictive models have been used in different areas of everyday life for decades. However, with the recent …
Koda Projekt DOI

National and regional hospitalization rates for allergic disorders in the United States : a 17 year time-trend analysis

Background: There is a paucity of data on national trends in hospitalizations for allergic disorders. The aim of this study was to …
PDF Projekt DOI

R you ready? Using the R programme for statistical analysis and graphics

For conducting research, nurses typically use commercial statistical packages. R software is a free, powerful, and flexible …
Koda Projekt DOI