Multi-class classification of heart sounds using enhanced bidirectional short-term memory metho
Bibliographic record
Abstract
The importance of early detection programs in improving the nation's overall health cannot be overemphasized. Conventional phonocardiogram (PCG) signals have been used a lot in this field of research because they are cheap and easy to use. The proposed convolutional neural network-based model differentiates the four types of cardiac auscultation just by looking at raw PCG data. To fully automate the process, two separate learning phases had to be created and given names that explained what they did. These learning stages are known as "representation learning" and "sequence residual learning.”. The use of this approach resulted in a significant enhancement in the precision of the PCG. This is because time has no effect on the qualities in any way. The findings of the study reveal that the suggested end-to-end architecture surpasses the most recent solutions in every area of the evaluation while utilizing nearly the same degree of computational complexity. Even though the proposed architecture has the same amount of computational complexity, this is still true. The recall rate of 99.52% demonstrates that the F1 score, accuracy, and precision all exceed 99.60%. In a head-to-head comparison with other models created using the same datasets as the other studies and publications, our model clearly wins. According to the study's findings, the model worked 85.57 percent of the time. The model worked and had an accuracy of 88.09 percent when tested with PhysioNet and Github PCG data. Extensive testing for primary and secondary validation yielded reliable results.
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 | 0.001 | 0.001 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 0.002 |
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".