The generalized sigmoid activation function: Competitive supervised learning

Narayan, Sridhar

doi:10.1016/s0020-0255(96)00200-9

Cited by 149 publications

(60 citation statements)

References 1 publication

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The MLP units take a number of real-valued inputs and generate a single real-valued output, according to an activation function (transfer function) applied to the weighted sum of the outputs of the units in the preceding layer. The most commonly used activation function in this network is a sigmoid function 40. The learning algorithm can be expressed using generalized Delta rule and back propagation gradient descent 41.…”

Section: Methodsmentioning

confidence: 99%

Profiles and Majority Voting-Based Ensemble Method for Protein Secondary Structure Prediction

Bouziane

Messabih

Chouarfia

2011

Evol Bioinform Online

View full text Add to dashboard Cite

Machine learning techniques have been widely applied to solve the problem of predicting protein secondary structure from the amino acid sequence. They have gained substantial success in this research area. Many methods have been used including k-Nearest Neighbors (k-NNs), Hidden Markov Models (HMMs), Artificial Neural Networks (ANNs) and Support Vector Machines (SVMs), which have attracted attention recently. Today, the main goal remains to improve the prediction quality of the secondary structure elements. The prediction accuracy has been continuously improved over the years, especially by using hybrid or ensemble methods and incorporating evolutionary information in the form of profiles extracted from alignments of multiple homologous sequences. In this paper, we investigate how best to combine k-NNs, ANNs and Multi-class SVMs (M-SVMs) to improve secondary structure prediction of globular proteins. An ensemble method which combines the outputs of two feed-forward ANNs, k-NN and three M-SVM classifiers has been applied. Ensemble members are combined using two variants of majority voting rule. An heuristic based filter has also been applied to refine the prediction. To investigate how much improvement the general ensemble method can give rather than the individual classifiers that make up the ensemble, we have experimented with the proposed system on the two widely used benchmark datasets RS126 and CB513 using cross-validation tests by including PSI-BLAST position-specific scoring matrix (PSSM) profiles as inputs. The experimental results reveal that the proposed system yields significant performance gains when compared with the best individual classifier.

show abstract

Section: Methodsmentioning

confidence: 99%

Profiles and Majority Voting-Based Ensemble Method for Protein Secondary Structure Prediction

Bouziane

Messabih

Chouarfia

2011

Evol Bioinform Online

View full text Add to dashboard Cite

show abstract

“…Both contained nodes to help capture nonlinearity in the input data, and an output layer, which contained a node to represent a dependent variable (VTA occurrence) 19,20 . We used rectified linear unit (RELU) 21 activation functions for the hidden layers, and the sigmoid activation function 22 for the output. The two hidden layers consisted of 22 neurons each.…”

Section: Sdmentioning

confidence: 99%

Application of a convolutional neural network for predicting the occurrence of ventricular tachyarrhythmia using heart rate variability features

Taye

Hwang

Lim

2020

Sci Rep

View full text Add to dashboard Cite

predicting the occurrence of ventricular tachyarrhythmia (VtA) in advance is a matter of utmost importance for saving the lives of cardiac arrhythmia patients. Machine learning algorithms have been used to predict the occurrence of imminent VtA. in this study, we used a one-dimensional convolutional neural network (1-D CNN) to extract features from heart rate variability (HRV), thereby to predict the onset of VtA. We also compared the prediction performance of our cnn with other machine leaning (ML) algorithms such as an artificial neural network (ANN), a support vector machine (SVM), and a k-nearest neighbor (KNN), which used 11 HRV features extracted using traditional methods. The proposed CNN achieved relatively higher prediction accuracy of 84.6%, while the ANN, SVM, and KNN algorithms obtained prediction accuracies of 73.5%, 67.9%, and 65.9% using 11 HRV features, respectively. Our result showed that the proposed 1-D CNN could improve VTA prediction accuracy by integrating the data cleaning, preprocessing, feature extraction, and prediction. Heartbeat is regulated by electrical signals conducted across the four chambers of the heart: two atria and two ventricles. When electrical activity is normal, the heart beats approximately 60 to 100 times per minute. However, abnormal electrical signals in the heart lead to disorganized electrical activities such as ventricular tachyarrhythmia (VTA), which causes fast heart rate 1. Thus, early VTA prediction helps physicians to take immediate medical procedure to reduce the risk. Ventricular tachycardia (VT) and ventricular fibrillation (VF) are the most common VTAs. VT arises from improper electrical activity in the ventricles, and can cause sudden cardiac arrest. VF is caused by chaotic electrical activity in the ventricles, which is similar to VT, but is a fatal condition that requires immediate medical attention. In VF, the heart shivers instead of pumping blood. Developing earlier preventive interventions would reduce the risk of experiencing an imminent VT and VF events. Researchers used noninvasive tests by measuring and analyzing electrocardiograms (ECGs), where heart rate variability (HRV) is extracted to train machine learning (ML) algorithms for predicting VT or VF in advance 2. HRV is the most commonly employed biomarker for isolating VT or VF subject from the normal subject 3. It is a time variation of heartbeats among two successive QRS complexes (Q, R, and S waves in ECG). In recent years, HRV indices have been used as a noninvasive biomarkers to forecast life-threatening arrhythmias 4. Previous studies mainly used the three traditional analysis methods: time domain, frequency domain, and Poincare nonlinear analyses, to extract features from HRV. Furthermore, they used these features as input to machine learning algorithms to predict the occurrence of VT, VF, or both. The machine learning techniques are used to classify the complex feature patterns and enable early prediction of VT or VF events with high accuracy. Acharya et al. used features extracted fr...

show abstract

“…Hence, this layer enables seamless interaction between different convolutional groups that extract patterns from different representations and facilitates joint feature learning from multiple information sources during back-propagation [44]. These complex non-linear features are then passed as inputs to a dense layer with softmax activation function [45], which draws a linear decision boundary on the derived feature space for separating the anticancer peptides from peptides without anticancer activity. Figure 1 represents the architecture of our proposed model for joint feature extraction from multiple information sources.…”

Section: Multi-headed Convolutional Neural Network Architecturementioning

confidence: 99%

ACP-MHCNN: An Accurate Multi-Headed Deep-Convolutional Neural Network to Predict Anticancer peptides

Ahmed

Muhammod

Adilina

et al. 2020

Preprint

View full text Add to dashboard Cite

Although advancing the therapeutic alternatives for treating deadly cancers has gained much attention globally, still the primary methods such as chemotherapy have significant downsides and low specificity. Most recently, Anticancer peptides (ACPs) have emerged as a potential alternative to therapeutic alternatives with much fewer negative side-effects. However, the identification of ACPs through wet-lab experiments is expensive and time-consuming. Hence, computational methods have emerged as viable alternatives. During the past few years, several computational ACP identification techniques using hand-engineered features have been proposed to solve this problem. In this study, we propose a new multi headed deep convolutional neural network model called ACP-MHCNN, for extracting and combining discriminative features from different information sources in an interactive way. Our model extracts sequence, physicochemical, and evolutionary based features for ACP identification through simultaneous interaction with different numerical peptide representations while restraining parameter overhead. It is evident through rigorous experiments using cross-validation and independent-dataset that ACP-MHCNN outperforms other models for anticancer peptide identification by a substantial margin. ACP-MHCNN outperforms state-of-the-art model by 6.3%, 8.6%, 3.7%, 4.0%, and 0.20 in terms of accuracy, sensitivity, specificity, precision, and MCC respectively. ACP-MHCNN and its relevant codes and datasets are publicly available at: https://github.com/mrzResearchArena/Anticancer-Peptides-CNN.

show abstract

The generalized sigmoid activation function: Competitive supervised learning

Cited by 149 publications

References 1 publication

Profiles and Majority Voting-Based Ensemble Method for Protein Secondary Structure Prediction

Profiles and Majority Voting-Based Ensemble Method for Protein Secondary Structure Prediction

Application of a convolutional neural network for predicting the occurrence of ventricular tachyarrhythmia using heart rate variability features

ACP-MHCNN: An Accurate Multi-Headed Deep-Convolutional Neural Network to Predict Anticancer peptides

Contact Info

Product

Resources

About