XGBoost machine learning algorithm for differential diagnosis of pediatric syncope
Bibliographic record
Abstract
Abstract. The search for new methods of differential diagnosis of syncope types will allow to improve the diagnosis of vasovagal syncope (VVS), syncope due to orthostatic hypotension (OH) and cardiac syncope (CS) in childhood in order to make timely adequate diagnostic and therapeutic decisions. The aim of the study was to develop an effective machine learning model for the differential diagnosis of VVS, syncope due to OH and CS in children. Materials and Methods. 140 patients with syncope, aged 8-17 years, were examined: 92 children with a diagnosis of VVS, 28 children with syncope due to OH and 20 children with CS. A machine learning model was built using XGBoost algorithm for multiclass classification based on input clinical, laboratory and instrumental patient data. Results. The developed machine learning model based on the XGBoost algorithm is effective in the differential diagnosis of VVS, syncope due to OH and CS, which is confirmed by the metrics of accuracy (0.93), precision (0.93 for VVS; 1.00 for syncope due to OH; 0.80 for CS), recall (0.96 for VVS; 1.00 for syncope due to OH; 0.67 for CS), f1 (0.95 for VVS; 1.00 for syncope due to OH; 0.73 for CS), ROC AUC (0.95 for VVS; 1.00 for syncope due to OH; 0.89 for CS), PR AUC (0.96 for VVS; 1.00 for syncope due to OH; 0.79 for CS),Cohen’s Kappa (0.85), and Matthews correlation coefficient (0.85). The most informative parameters of the syncope types differential diagnosis model are OH, paroxysmal supraventricular tachycardia, Hildebrandt coefficient, Calgary Syncope Seizure Score, vitamin B6, average duration of the P-Q interval during 24 hours, duration of tachycardia during 24 hours, stroke index, homocysteine, heart volume, and systolic blood volume. Conclusions. The proposed machine learning model has sufficient efficiency and can be used by pediatricians and pediatric cardiologists for the differential diagnosis of VS, syncope due to OH, and CS in childhood.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".