Contrastive learning of heart and lung sounds for label-efficient diagnosis

Soni, Pratham N.; Shi, Siyu; Sriram, Pranav R.; Ng, Andrew Y.; Rajpurkar, Pranav

doi:10.1016/j.patter.2021.100400

Cited by 23 publications

(10 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The performance of the proposed methodology was the highest in the experimental results, and the effects of attention were confirmed. Moreover, Soni et al [53] proposed a contrastive learning classification methodology using the dataset as an external validation set. Data was useful enough to be used for model validation.…”

Section: Resultsmentioning

confidence: 99%

Interpretation of lung disease classification with light attention connected module

Choi

Lee

2023

Biomedical Signal Processing and Control

View full text Add to dashboard Cite

Section: Resultsmentioning

confidence: 99%

Interpretation of lung disease classification with light attention connected module

Choi

Lee

2023

Biomedical Signal Processing and Control

View full text Add to dashboard Cite

“…Numerous businesses, including banking and online businesses, utilize this calculation to foresee client conduct and results. A few choice trees make up an irregular woodland calculation [26]. Bagging or bootstrap conglomeration is utilized to prepare the "woodland" of the arbitrary backwoods calculation [26].…”

Section: Random Forestmentioning

confidence: 99%

“…A few choice trees make up an irregular woodland calculation [26]. Bagging or bootstrap conglomeration is utilized to prepare the "woodland" of the arbitrary backwoods calculation [26]. A meta-algorithm group known as "bagging" is utilized to work on the precision of the AI algorithm of diferent sorts.…”

Section: Random Forestmentioning

confidence: 99%

Application of Machine Learning for Cardiovascular Disease Risk Prediction

Dalal

Goel

Onyema

et al. 2023

Computational Intelligence and Neuroscience

View full text Add to dashboard Cite

Cardiovascular diseases (CVDs) are a common cause of heart failure globally. The need to explore possible ways to tackle the disease necessitated this study. The study designed a machine learning model for cardiovascular disease risk prediction in accordance with a dataset that contains 11 features which may be used to forecast the disease. The dataset from Kaggle on cardiovascular disease includes approximately 70,000 patient records that were used to determine the outcome. Compared to the UCI dataset, the Kaggle dataset has many more training and validation records. Models created using neural networks, random forests, Bayesian networks, C5.0, and QUEST were compared for this dataset. On training and testing data sets, the results acquired a high accuracy (99.1 percent), which is significantly superior to previous methods. Ahead-of-time detection and diagnosis of cardiac disease, as well as better treatment outcomes, are strong possibilities for the suggested prediction model. Additionally, it may help patients better manage their illness or life forms in order to increase their chances of recovery/survival. The result showed greater accuracy and promising signs that machine-learning algorithms can indeed assist in early identification of the disease and improvement of the treatment outcome.

show abstract

“…Self‐supervised learning of AI has shown potential in addressing the shortage of labeled data by allowing algorithms to train with unlabeled data 13 . Early results of this technique show that it can reduce the amount of labeled data required for the development of medical imaging algorithms 14 . Also, systems are being developed to create simulated data 15 .…”

Section: Figurementioning

confidence: 99%