Comparison of Unsupervised Machine Learning Approaches for Cluster Analysis to Define Subgroups of Heart Failure with Preserved Ejection Fraction with Different Outcomes
Bibliographic record
Abstract
Heart failure with preserved ejection (HFpEF) is a heterogenous condition affecting nearly half of all patients with heart failure (HF). Artificial intelligence methodologies can be useful to identify patient subclassifications with important clinical implications. We sought a comparison of different machine learning (ML) techniques and clustering capabilities in defining meaningful subsets of patients with HFpEF. Three unsupervised clustering strategies, hierarchical clustering, K-prototype, and partitioning around medoids (PAM), were used to identify distinct clusters in patients with HFpEF, based on a wide range of demographic, laboratory, and clinical parameters. The study population had a median age of 77 years, with a female majority, and moderate diastolic dysfunction. Hierarchical clustering produced six groups but two were too small (two and seven cases) to be clinically meaningful. The K-prototype methods produced clusters in which several clinical and biochemical features did not show statistically significant differences and there was significant overlap between the clusters. The PAM methodology provided the best group separations and identified six mutually exclusive groups (HFpEF1-6) with statistically significant differences in patient characteristics and outcomes. Comparison of three different unsupervised ML clustering strategies, hierarchical clustering, K-prototype, and partitioning around medoids (PAM), was performed on a mixed dataset of patients with HFpEF containing clinical and numerical data. The PAM method identified six distinct subsets of patients with HFpEF with different long-term outcomes or mortality. By comparison, the two other clustering algorithms, the hierarchical clustering and K-prototype, were less optimal.
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 machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.012 | 0.027 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.002 |
| Bibliometrics | 0.004 | 0.002 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.002 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 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 source (direct Gemma or distilled Codex), 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".