Short term prediction of atmospheric temperature using neural networks

In this paper, a neural network based forecasting model for the maximum and the minimum temperature for the ground level is proposed. A backpropagation method of gradient-decent learning in multi-layer perceptron (MLP) type of neural network with only one hidden layer is considered. This network consists of 25 input nodes and two output nodes. The network is trained with a varying number of nodes in the hidden layer using a set of training sample and each of them is tested with a set of test sample. It accepts previous two consecutive days information (such as pressures, temperatures, relative humidities, etc.) as inputs for the estimation of the maximum and the minimum temperature as output. The network with 20 or less neurons in the hidden layer is found to be "optimum" and it produces an error within ±2° C for 80% of test cases.

Assessing Feature Importance for Short-Term Prediction of Electricity Demand in Medium-Voltage Loads

The design of new monitoring systems for intelligent distribution networks often requires both real-time measurements and pseudomeasurements to be processed. The former are obtained from smart meters, phasor measurement units and smart electronic devices, whereas the latter are predicted using appropriate algorithms—with the typical objective of forecasting the behaviour of power loads and generators. However, depending on the technique used for data encoding, the attempt at making predictions over a period of several days may trigger problems related to the high number of features. To contrast this issue, feature importance analysis becomes a tool of primary importance. This article is aimed at illustrating a technique devised to investigate the importance of features on data deemed relevant for predicting the next hour demand of aggregated, medium-voltage electrical loads. The same technique allows us to inspect the hidden layers of multilayer perceptrons entrusted with making the predictions, since, ultimately, the content of any hidden layer can be seen as an alternative encoding of the input data. The possibility of inspecting hidden layers can give wide support to researchers in a number of relevant tasks, including the appraisal of the generalisation capability reached by a multilayer perceptron and the identification of neurons not relevant for the prediction task.

Detection of IoT based DDoS Attacks by Network Traffic Analysis using Feedforward Neural Networks

In this paper an optimized feedforward neural network model is proposed for detection of IoT based DDoS attacks by network traffic analysis aimed towards a specific target which could be constantly monitored by a tap. The proposed model is applicable for DoS and DDoS attacks which consist of TCP, UDP and HTTP flood and also against keylogging, data exfiltration, OS fingerprint and service scan activities. It simply differentiates such kind of network traffic from normal network flows. The neural network uses Adam optimization as a solver and the hyperbolic tangent activation function in all neurons from a single hidden layer. The number of hidden neurons could be varied, depending on targeted accuracy and processing speed. Testing over the Bot IoT dataset reveals that developed models are applicable using 8 or 10 features and achieved discrimination error of 4.91.10-3%.

Dual Co-Attention-Based Multi-Feature Fusion Method for Rumor Detection

Social media has become more popular these days due to widely used instant messaging. Nevertheless, rumor propagation on social media has become an increasingly important issue. The purpose of this study is to investigate the impact of various features in social media on rumor detection, propose a dual co-attention-based multi-feature fusion method for rumor detection, and explore the detection capability of the proposed method in early rumor detection tasks. The proposed BERT-based Dual Co-attention Neural Network (BDCoNN) method for rumor detection, which uses BERT for word embedding . It simultaneously integrates features from three sources: publishing user profiles, source tweets, and comments. In the BDCoNN method, user discrete features and identity descriptors in user profiles are extracted using a one-dimensional convolutional neural network (CNN) and TextCNN, respectively. The bidirectional gate recurrent unit network (BiGRU) with a hierarchical attention mechanism is used to learn the hidden layer representation of tweet sequence and comment sequence. A dual collaborative attention mechanism is used to explore the correlation among publishing user profiles, tweet content, and comments. Then the feature vector is fed into classifier to identify the implicit differences between rumor spreaders and non-rumor spreaders. In this study, we conducted several experiments on the Weibo and CED datasets collected from microblog. The results show that the proposed method achieves the state-of-the-art performance compared with baseline methods, which is 5.2% and 5% higher than the dEFEND. The F1 value is increased by 4.4% and 4%, respectively. In addition, this paper conducts research on early rumor detection tasks, which verifies the proposed method detects rumors more quickly and accurately than competitors.

Prototyping Feed-Forward Artificial Neural Network on Spartan 3S1000 FPGA for Blood Type Classification

In this research, a Feed-Forward Artificial Neural Network design was implemented on Xilinx Spartan 3S1000 Field Programable Gate Array using XSA-3S Board and prototyped blood type classification device. This research uses blood sample images as a system input. The system was built using VHSIC Hardware Description Language to describe the feed-forward propagation with a backpropagation neural network algorithm. We use three layers for the feed-forward ANN design with two hidden layers. The hidden layer designed has two neurons. In this study, the accuracy of detection obtained for four-type blood image resolutions results from 86%-92%, respectively.

Drought prediction based on an improved VMD-OS-QR-ELM model

To overcome the low accuracy, poor reliability, and delay in the current drought prediction models, we propose a new extreme learning machine (ELM) based on an improved variational mode decomposition (VMD). The model first redefines the output of the hidden layer of the ELM model with orthogonal triangular matrix decomposition (QR) to construct an orthogonal triangular ELM (QR-ELM), and then introduces an online sequence learning mechanism (OS) into the QR-ELM to construct an online sequence OR-ELM (OS-QR-ELM), which effectively improves the efficiency of the ELM model. The mutual information extension method was then used to extend both ends of the original signal to improve the VMD end effect. Finally, VMD and OS-QR-ELM were combined to construct a drought prediction method based on the VMD-OS-QR-ELM. The reliability and accuracy of the VMD-OS-QR-ELM model were improved by 86.19% and 93.20%, respectively, compared with those of the support vector regression model combined with empirical mode decomposition. Furthermore, the calculation efficiency of the OS-QR-ELM model was increased by 88.65% and 85.32% compared with that of the ELM and QR-ELM models, respectively.

Prediction of Bacteremia Based on 12-Year Medical Data Using a Machine Learning Approach: Effect of Medical Data by Extraction Time

Early detection of bacteremia is important to prevent antibiotic abuse. Therefore, we aimed to develop a clinically applicable bacteremia prediction model using machine learning technology. Data from two tertiary medical centers’ electronic medical records during a 12-year-period were extracted. Multi-layer perceptron (MLP), random forest, and gradient boosting algorithms were applied for machine learning analysis. Clinical data within 12 and 24 hours of blood culture were analyzed and compared. Out of 622,771 blood cultures, 38,752 episodes of bacteremia were identified. In MLP with 128 hidden layer nodes, the area under the receiver operating characteristic curve (AUROC) of the prediction performance in 12- and 24-h data models was 0.762 (95% confidence interval (CI); 0.7617–0.7623) and 0.753 (95% CI; 0.7520–0.7529), respectively. AUROC of causative-pathogen subgroup analysis predictive value for Acinetobacter baumannii bacteremia was the highest at 0.839 (95% CI; 0.8388–0.8394). Compared to primary bacteremia, AUROC of sepsis caused by pneumonia was highest. Predictive performance of bacteremia was superior in younger age groups. Bacteremia prediction using machine learning technology appeared possible for acute infectious diseases. This model was more suitable especially to pneumonia caused by Acinetobacter baumannii. From the 24-h blood culture data, bacteremia was predictable by substituting only the continuously variable values.

Particle Swarm Optimization of BP-ANN Based Soft Sensor for Greenhouse Climate

In this article, the authors develop the Particle Swarm Optimization algorithm (PSO) in order to optimise the BP network in order to elaborate an accurate dynamic model that can describe the behavior of the temperature and the relative humidity under an experimental greenhouse system. The PSO algorithm is applied to the Back-Propagation Neural Network (BP-NN) in the training phase to search optimal weights baded on neural networks. This approach consists of minimising the reel function which is the mean squared difference between the real measured values of the outputs of the model and the values estimated by the elaborated neural network model. In order to select the model which possess higher generalization ability, various models of different complexity are examined by the test-error procedure. The best performance is produced by the usage of one hidden layer with fourteen nodes. A comparison of measured and simulated data regarding the generalization ability of the trained BP-NN model for both temperature and relative humidity under greenhouse have been performed and showed that the elaborated model was able to identify the inside greenhouse temperature and humidity with a good accurately.

Evaluation of Parameter Settings for Training Neural Networks Using Backpropagation Algorithms

Artificial neural networks (ANN) are widely used for classification, and the training algorithm commonly used is the backpropagation (BP) algorithm. The major bottleneck faced in the backpropagation neural network training is in fixing the appropriate values for network parameters. The network parameters are initial weights, biases, activation function, number of hidden layers and the number of neurons per hidden layer, number of training epochs, learning rate, minimum error, and momentum term for the classification task. The objective of this work is to investigate the performance of 12 different BP algorithms with the impact of variations in network parameter values for the neural network training. The algorithms were evaluated with different training and testing samples taken from the three benchmark clinical datasets, namely, Pima Indian Diabetes (PID), Hepatitis, and Wisconsin Breast Cancer (WBC) dataset obtained from the University of California Irvine (UCI) machine learning repository.