Nätverksanalys och Maskininlärning av Triagedata inom BUP Region Skåne 2010-2023
(2026) EEML05 20261Division for Biomedical Engineering
- Abstract
- An increasing challenge in modern society is the rising prevalence of psychiatric problems among children and adolescents. Child and Adolescent Psychiatry (BUP) works with triage, treatment and follow-up of young patients, generating large amounts of clinical data throughout the treatment process. Despite this, the clinical data are currently limited used. The aim of this project was therefore to use machine learning, partial correlation network analysis and statistical methods such as descriptive and clinical cut-off analysis to identify patterns andlong-term trends in BUP patient data.
Longitudinal triage data collected between 2010 and 2023 were cleaned and analyzed to investigate how children’s symptoms and functioning have changed... (More) - An increasing challenge in modern society is the rising prevalence of psychiatric problems among children and adolescents. Child and Adolescent Psychiatry (BUP) works with triage, treatment and follow-up of young patients, generating large amounts of clinical data throughout the treatment process. Despite this, the clinical data are currently limited used. The aim of this project was therefore to use machine learning, partial correlation network analysis and statistical methods such as descriptive and clinical cut-off analysis to identify patterns andlong-term trends in BUP patient data.
Longitudinal triage data collected between 2010 and 2023 were cleaned and analyzed to investigate how children’s symptoms and functioning have changed over time, as well as to evaluate whether machine learning methods could identify more complex patterns by predicting year and two functioning variables using Regression Learner in MATLAB.
The statistical analyses identified several trends over time, including worsening symptom and functional profiles along withdecreasing patient age.
However, the machine learning models showed limited ability to predict year based on symptom and functional variables, suggesting that changes over time are gradual and influencedby multiple interacting factors rather than year-specific patterns. The prediction of functioning variables was also relatively weak. Yet, the analyses indicated a stronger association between psychiatric symptoms and school functioning compared with family functioning.
These findings suggest that machine learning can serve as a complementary analysis method for longitudinal psychiatric data. However, overall changes in psychiatric symptoms over time were more clearly identified using basic statistical methods and visualization techniques. (Less)
Please use this url to cite or link to this publication:
https://lup.lub.lu.se/student-papers/record/9236577
- author
- Elmstrand, Nora LU and Bergkvist, Amanda LU
- supervisor
- organization
- alternative title
- Network Analysis and Machine Learning of Triage Data within BUP Region Skåne 2010-2023
- course
- EEML05 20261
- year
- 2026
- type
- M2 - Bachelor Degree
- subject
- keywords
- Maskininlärning, psykisk ohälsa, BUP, psykiatri
- language
- Swedish
- id
- 9236577
- date added to LUP
- 2026-06-23 12:42:47
- date last changed
- 2026-06-23 12:42:47
@misc{9236577,
abstract = {{An increasing challenge in modern society is the rising prevalence of psychiatric problems among children and adolescents. Child and Adolescent Psychiatry (BUP) works with triage, treatment and follow-up of young patients, generating large amounts of clinical data throughout the treatment process. Despite this, the clinical data are currently limited used. The aim of this project was therefore to use machine learning, partial correlation network analysis and statistical methods such as descriptive and clinical cut-off analysis to identify patterns andlong-term trends in BUP patient data.
Longitudinal triage data collected between 2010 and 2023 were cleaned and analyzed to investigate how children’s symptoms and functioning have changed over time, as well as to evaluate whether machine learning methods could identify more complex patterns by predicting year and two functioning variables using Regression Learner in MATLAB.
The statistical analyses identified several trends over time, including worsening symptom and functional profiles along withdecreasing patient age.
However, the machine learning models showed limited ability to predict year based on symptom and functional variables, suggesting that changes over time are gradual and influencedby multiple interacting factors rather than year-specific patterns. The prediction of functioning variables was also relatively weak. Yet, the analyses indicated a stronger association between psychiatric symptoms and school functioning compared with family functioning.
These findings suggest that machine learning can serve as a complementary analysis method for longitudinal psychiatric data. However, overall changes in psychiatric symptoms over time were more clearly identified using basic statistical methods and visualization techniques.}},
author = {{Elmstrand, Nora and Bergkvist, Amanda}},
language = {{swe}},
note = {{Student Paper}},
title = {{Nätverksanalys och Maskininlärning av Triagedata inom BUP Region Skåne 2010-2023}},
year = {{2026}},
}