A New Neural Network with Adaptive Activation Function for Classification of ECG Arrhythmias

Ectopic beats detection and correction methods: A review

“…ANN can be used in pattern recognition or data classification. Neural networks were successfully applied in the arrhythmia classification [45,46]. In fact, the choice of the input features plays an important role in the classification or pattern recognition tasks.…”

A new method for classification of ECG arrhythmias using neural network with adaptive activation function

“…In a previous study of ours [34], various architectures and configurations for NNAAF-1 and NNAAF-2 have been tried and tested with records from 10 patients. Differing from that, in present study, architectures of NNAAF-1 and NNAAF-2 have been altered to obtain smaller error values.…”

Towards an Intelligent Framework for Pressure-Based 3D Curve Drawing

“…This means a properly trained NN is able to correctly classify data which is out of the training dataset. Due to its unique capability to generalize in the presence of noise, NN has been used widely in signal or waveform analysis in the field of medical [16][17][18][19][20][21][22][23][24][25] and finance [26]. Neural network has been explored in the financial market for Elliot waveform recognition [26] to identify and predict repeating pattern in future trends.…”

“…They are representative signals that contain valuable information to the nature of the disease, which is reflected in the shape of the waveform [16,17,[20][21][22][23]25]. The common characteristics of these bio-signals are that they are non stationary, contaminated with noise, and have large variation in the morphologies of the waveform not only of different patients or patient groups but also within the same patient [17,21,22,25]. Numerous other waveform classifier such as beat classifier, digital filter, linear and non linear methods have been explored previously, all perform well on training data but generalize poorly [21,22,24,25].…”

“…The unique ability of neural network to generalize the variation of waveform morphology has made neural network a preferred and reliable waveform classifier. A great variety of neural networks have been experimented in previous work for effective classification of ECG [17,[20][21][22][23][24][25], EEG [16], Doppler [18,19], and Elliot waveform [26]. All reported good recognition rate with correct classification in over 90% of the cases.…”