Multi-classification neural network model for detection of abnormal heartbeat audio signals

Malik, Hassaan; Bashir, Umair; Ahmad, Adnan

doi:10.1016/j.bea.2022.100048

Cited by 14 publications

(10 citation statements)

References 44 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It achieved an accuracy of 77.5% and a sensitivity of 75.2%, both of which are higher than those of the other two experiments. According to the results of the performance study of RF, the MLP model is superior to various other models [65][66][67][68][69][70][71][72]. We performed a 15fold LOO to further test the outcomes of the proposed model, and the result validates the relevance of MLP, as can be shown in Table 3.…”

Section: Discussionmentioning

confidence: 79%

“…The window is considered positive if the middle of the window binds the protein nucleotide with (+) instance, and the window containing non-binding nucleotide is considered negative. We also removed the feature vectors that were neither (+) nor (-) from the training phase because it may produce severely unbalanced instances for training, unless and until the initial and final windows were supposed as (+) if they consist of a protein-binding nucleotide in any location of the sliding window [68][69][70][71][72]. This is due to the limited number of (+) instances than (-) in the training database.…”

Section: Endmentioning

confidence: 99%

See 1 more Smart Citation

Neural Network-Based Prediction of Potential Ribonucleic Acid Aptamers to Target Protein

Mahmood¹,

Malik²,

Khan³

et al. 2022

JCBI

View full text Add to dashboard Cite

Aptamers are short strands of nucleic acid with a single strand that may unite to target a certain molecule in a selective and specific manner. SELEX experiments are the typical method used for identifying aptamers in vitro (systematic evolution of ligands by exponential enrichment). Several different computational methods have been developed to locate aptamers. The purpose of this research is to identify and make predictions on the possible RNA aptamers that may be used to target the protein. To do this, we propose the use of a multi-layer perceptron neural network with sixteen layers that are trained to locate possible aptamers of a protein target. This network is trained by extracting the main properties of RNA sequences. The outcome of our proposed model is compared to the output of two well-known machine learning classifiers, namely random forest (RF) and support vector machine (SVM). Additionally, we undertake the independent testing of our model on the benchmark dataset, which allows us to reach the highest accuracy possible. As a consequence of this, our model obtains an accuracy of 98.44% and an MCC of 0.9123 during the 15-fold cross-validation, and it achieves an accuracy of 98.10% and an MCC of 0.9354 when the leave-one-out cross-validation is performed. We are certain that our approach will contribute to a reduction in the amount of money and time spent on in vitro testing. Therefore, restricting the length of the initial pool of potential nucleic acid pattern combinations.

show abstract

Section: Discussionmentioning

confidence: 79%

Section: Endmentioning

confidence: 99%

Neural Network-Based Prediction of Potential Ribonucleic Acid Aptamers to Target Protein

Mahmood¹,

Malik²,

Khan³

et al. 2022

JCBI

View full text Add to dashboard Cite

show abstract

“…Deep neural networks [3, were a part of the pre-trained image classifier; however, the final convolutional layers of these networks resulted in a loss of the spatial resolution of the feature maps, which in turn restricted the classification capabilities of these models. In addition, the large number of neurons coupled to the input resulted in an insufficiently large filter size for these networks [51][52][53][54][55][56][57][58][59][60][61][62][63][64][65][66][67][68][69][70]73,74]. Because of this, the network could overlook significant traits as soon as they were discovered, which is a problem in how the network was designed.…”

Section: Discussionmentioning

confidence: 99%

Deep Learning-Based Classification of Abrasion and Ischemic Diabetic Foot Sores Using Camera-Captured Images

Khalil,

Naeem,

Naqvi

et al. 2023

Mathematics

View full text Add to dashboard Cite

Diabetic foot sores (DFS) are serious diabetic complications. The patient’s weakened neurological system damages the tissues of the foot’s skin, which results in amputation. This study aims to validate and deploy a deep learning-based system for the automatic classification of abrasion foot sores (AFS) and ischemic diabetic foot sores (DFS). We proposed a novel model combining convolutional neural network (CNN) capabilities with Vgg-19. The proposed method utilized two benchmark datasets to classify AFS and DFS from the patient’s foot. A data augmentation technique was used to enhance the accuracy of the training. Moreover, image segmentation was performed using UNet++. We tested and evaluated the proposed model’s classification performance against two well-known pre-trained classifiers, Inceptionv3 and MobileNet. The proposed model classified AFS and ischemia DFS images with an accuracy of 99.05%, precision of 98.99%, recall of 99.01%, MCC of 0.9801, and f1 score of 99.04%. Furthermore, the results of statistical evaluations using ANOVA and Friedman tests revealed that the proposed model exhibited a remarkable performance. The proposed model achieved an excellent performance that assist medical professionals in identifying foot ulcers.

show abstract

“…The study described in [16] introduced an RNN model that used LSTM. The model was trained using two widely accessible datasets: the PASCAL and the 2017 PhysioNet challenges.…”

Section: Related Work On Hybrid Deep Learning Model For Multiclass Ar...mentioning

confidence: 99%

A hybrid deep learning model for multiclass arrhythmia detection and classification using photoplethysmography signals

Kumar,

2024

Preprint

View full text Add to dashboard Cite

Arrhythmias (AM) are a typical kind of heart ailment that often directly leads to fatal cardiac arrest. Therefore, the automated identification of arrhythmias is crucial in aiding the detection of cardiac diseases. Previous studies have documented the use of photoplethysmography (PPG) signals to identify arrhythmias. However, there is little research on the application of PPG signals for the classification of multiclass arrhythmias. This research aims to determine whether PPG signals and a Hybrid Deep Learning (HDL) network could effectively categorize various forms of arrhythmia in a multiclass setting. The purpose of this study is to introduce an HDL model called DenseNet-ABiLSTM, which is specifically developed for the identification and categorization of various types of arrhythmias. This technique leverages the densely connected convolutional network's multiscale spatial feature extraction capability (DenseNet) and the benefit of Attention-based Bidirectional Long Short-Term Memory (ABiLSTM) with enhanced attention mechanism in sequence modelling. The PPG signal is initially subjected to 1D convolutional kernels to acquire multiscale conceptual features. Subsequently, the BiLSTM is used to get insight into the temporal relationship among the features. The Attention Mechanism (AtM) layer is presented to emphasize its significant features for achieving improved detection performance. The characteristic data is ultimately related to the associated arrhythmia rhythms classified using the SoftMax layer. The model categorizes arrhythmia rhythms into six types: Sinus Rhythm (SR), Early Ventricular Contraction (EVC), Early Atrial Contraction (EAC), Ventricular Tachycardia (VcT), Supraventricular Tachycardia (ST), and AF. Various metrics have assessed the classification performance and have been compared with the Electrocardiogram (ECG) results to determine the AM rhythms. Two hundred twenty-five patients and 109736 pairs of 12-second PPG and ECG waveforms were utilized. The mean performance measures show strong overall performance despite minor fluctuations across various rhythms, with a mean F1 score and accuracy of 87.74% and 89.14%, respectively.

show abstract

Multi-classification neural network model for detection of abnormal heartbeat audio signals

Cited by 14 publications

References 44 publications

Neural Network-Based Prediction of Potential Ribonucleic Acid Aptamers to Target Protein

Neural Network-Based Prediction of Potential Ribonucleic Acid Aptamers to Target Protein

Deep Learning-Based Classification of Abrasion and Ischemic Diabetic Foot Sores Using Camera-Captured Images

A hybrid deep learning model for multiclass arrhythmia detection and classification using photoplethysmography signals

Contact Info

Product

Resources

About