Search

A multi-layer convolutional neural network (MCNN) with hyperparameter optimization (HyperMCNN) is proposed for classifying human electrocardiograms (ECGs). For performance tests of the HyperMCNN, ECG recordings for patients with cardiac arrhythmia (ARR), congestive heart failure (CHF), and normal sinus rhythm (NSR) were obtained from three PhysioNet databases: MIT-BIH Arrhythmia Database, BIDMC Congestive Heart Failure Database, and MIT-BIH Normal Sinus Rhythm Database, respectively. The MCNN hyperparameters in convolutional layers included number of filters, filter size, padding, and filter stride. The hyperparameters in max-pooling layers were pooling size and pooling stride. Gradient method was also a hyperparameter used to train the MCNN model. Uniform experimental design approach was used to optimize the hyperparameter combination for the MCNN. In performance tests, the resulting 16-layer CNN with an appropriate hyperparameter combination (16-layer HyperMCNN) was used to distinguish among ARR, CHF, and NSR. The experimental results showed that the average correct rate and standard deviation obtained by the 16-layer HyperMCNN were superior to those obtained by a 16-layer CNN with a hyperparameter combination given by Matlab examples. Furthermore, in terms of performance in distinguishing among ARR, CHF, and NSR, the 16-layer HyperMCNN was superior to the 25-layer AlexNet, which was the neural network that had the best image identification performance in the ImageNet Large Scale Visual Recognition Challenge in 2012.

A large categories of time series fluctuate dramatically, for example, prices of agriculture produce. Traditional methods in time series and stochastic prediction may not capture such dynamics. This paper tries to use machine learning to tune the model for a real situation by establishing a price determination mechanism on the model of stochastic automata (SA) and evolutionary game (EG). Time series volatility attributed to the chaotic process can be obtained through the learning algorithm of Markov Chain Monte Carlo (MCMC). Using machine learning through the chaotic analysis of stochastic automata and evolutionary games, we find that a more spatially aggregated distribution (smaller entropy) leads to larger time series fluctuations, regardless of the initial distribution of crops. By integrating the factors discovered in this study, we can develop a better learning algorithm in a highly volatile time series in agriculture prices.

The uncertainty issue in real-work optimization affects the level of optimization significantly. Because most future uncertainties cannot be foreseen in advance, the optimization must take the uncertainties as a risk in an intelligent way in the process of computation algorithm. Based on our risk-sensitive filtering algorithm, this study adopts a model-predictive control to construct a risk-averse, predictable model that can be used to regulate the level of a real-world system. Our model is intelligent in that the predictive model needs not to identify the system parameters in advance, and our algorithm will learn the parameters through data. When the real-world system is under the disturbance of unexpected events, our model can still maintain suitable performance. Our results show that the intelligent model designed in this study can learn the system parameters in a real-world system and minimize unexpected real-world disturbances. Through the learning process, our model is robust, and the optimal performance can still be retained even the system parameters deviate from expected, e.g., material shortage in a supply chain due to earthquake. When parameter error risks occur, the control rules can still drive the overall system with a minimal performance drop.

This study proposes a framework implementing triangular estimation for better modeling and forecasting time series. In order to improve the performance of estimation, we employ two sources of triangulation to generate a time series, which is statistically indistinguishable with the latent time series hidden in a system. Thanks to Bayesian hierarchical estimation, which is akin to deep learning but more sophisticate and longer history, the framework has been validated by a large amount of records in vegetable auctions. The hierarchical Bayesian estimation and Monte Carlo Markov Chain particle filters used in hidden Markov model are appreciated during the massive bootstrapping of data. Our results demonstrate excellent estimation performance in discovering hidden states.

Atrial fibrillation (AF) is a type of paroxysmal cardiac disease that presents no obvious symptoms during onset, and even the electrocardiograms (ECG) results of patients with AF appear normal under a premorbid status, rendering AF difficult to detect and diagnose. However, it can result in deterioration and increased risk of stroke if not detected and treated early. This study used the ECG database provided by the Physionet website (https://physionet.org), filtered data, and employed parameter-extraction methods to identify parameters that signify ECG features. A total of 31 parameters were obtained, consisting of P-wave morphology parameters and heart rate variability parameters, and the data were further examined by implementing a decision tree, of which the topmost node indicated a significant causal relationship. The experiment results verified that the P-wave morphology parameters significantly affected the ECG results of patients with AF.