Phonocardiogram Classification Model With Kolmogorov–Arnold Network for Training With Heterogeneous Dataset
Bibliographic record
Abstract
Phonocardiogram (PCG) can be used to detect cardiac conditions and support the initial diagnosis of cardiovascular disease, a critical health issue that requires early detection to allow timely treatment and potentially save lives. Classification of PCG signals as normal or abnormal is currently done using learning algorithms which require homogeneous training data. However, PCG datasets are often collected using stethoscopes with varying characteristics, from different individuals, and in diverse controlled or uncontrolled environments. This results in dataset heterogeneity, which poses a challenge for training effective deep learning models. This study explores the recently proposed Kolmogorov–Arnold Networks (KAN), which incorporate different trainable function families such as splines and wavelets for the classification of PCG and evaluate their robustness against data heterogeneity. KAN is compared with a traditional Multi-Layer Perceptron (MLP) on heterogeneous and homogeneous PCG datasets to determine the most suitable model for PCG classification. Experimental results show that KAN with wavelet-based functions outperforms KAN with spline functions and MLP on both datasets, achieving superior performance with parameters and computational costs comparable to those of MLP. In contrast, the spline-based KAN performs well on homogeneous data but poorly on heterogeneous data, incurring the highest computational cost and model complexity. KAN with wavelet functions outperforms MLP by over 10% in most cases and outperforms state-of-the art methods. In summary, KAN with wavelet functions demonstrate strong performance across dataset types and may be a promising candidate for fully connected layers in deep learning models, irrespective of whether the dataset is homogeneous or heterogeneous.
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) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.002 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.001 |
| 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".