查看更多>>摘要：Artificial intelligence technology has been widely used in various fields in recent years. In the case of typhoons, trajectory prediction technology can reduce the loss of human life and property caused by typhoon movements. From the perspective of deep learning, multimodal learning and multitask learning are applied to trajectory prediction. And a trajectory prediction model based on deep multimodal fusion and multitask generation (Trj-DMFMG) is proposed. The model mainly includes two modules: a deep multimodal fusion module and a multitask generation module. The deep multimodal fusion module is composed of several multimodal fusion modules. First, the multimodal trajectory sequence is divided into multiple multimodal subtrajectories by using a sliding window. Then, the multimodal fusion module trains different modal data to perform feature fusion through a long shortterm memory network (LSTM) and a 3D convolution neural network (3D CNN). Finally, the features generated by multiple multimode fusion modules are deeply fused. The multitask generation module first trains the deep fusion features generated by the deep multimodal fusion module through the LSTM, then it realizes longitude and latitude prediction at the same time. In this paper, real typhoon data in the Northwest Pacific Ocean are used for simulation experiments. Through a comprehensive comparison of the prediction results in longitude and latitude, it is found that Trj-DMFMG has the best prediction effect and is more accurate and stable in long-term prediction. (c) 2022 Elsevier B.V. All rights reserved.
查看更多>>摘要：COVID-19 pandemic caused by novel coronavirus (SARS-CoV-2) crippled the world economy and engendered irreparable damages to the lives and health of millions. To control the spread of the disease, it is important to make appropriate policy decisions at the right time. This can be facilitated by a robust mathematical model that can forecast the prevalence and incidence of COVID-19 with greater accuracy. This study presents an optimized ARIMA model to forecast COVID-19 cases. The proposed method first obtains a trend of the COVID-19 data using a low-pass Gaussian filter and then predicts/forecasts data using the ARIMA model. We benchmarked the optimized ARIMA model for 7-days and 14-days forecasting against five forecasting strategies used recently on the COVID-19 data. These include the auto-regressive integrated moving average (ARIMA) model, susceptible-infected-removed (SIR) model, composite Gaussian growth model, composite Logistic growth model, and dictionary learning-based model. We have considered the daily infected cases, cumulative death cases, and cumulative recovered cases of the COVID-19 data of the ten most affected countries in the world, including India, USA, UK, Russia, Brazil, Germany, France, Italy, Turkey, and Colombia. The proposed algorithm outperforms the existing models on the data of most of the countries considered in this study. (C)& nbsp;2022 Elsevier B.V. All rights reserved.
查看更多>>摘要：Multi-face detection and alignment techniques under unlimited environment are challenging issues. In recent years, the demand for face detection and alignment techniques has increased in many areas, including automatic drive and security. However, some mainstream algorithms, such as the Multi-task Cascaded Convolutional Networks (MTCNN) algorithm cannot transfer to multi-face detection and alignment, which introduces new multi-size, multi-resolution, and multi-angle challenges. This paper proposes an improved algorithm-Multi-face-MTCNN for precise original face detection and alignment algorithm when there is an overlapping face scenario. We design two new network structures: Pixelfusion-MTCNN and Twoconv-MTCNN. Moreover, we propose new data augmentation method and optimized detection process which are applied in the Multi-face-MTCNN algorithm. The limitations associated with single-scale kernel size in MTCNN are solved to obtain a satisfactory performance. Compared to the MTCNN algorithm, experimental analysis of the FDDB dataset show that 1.766% improves Multi-face-MTCNN. Meanwhile, on the WIDER FACE verification benchmark, respectively, with regards to the three sub-datasets, the proposed algorithm's performance is improved by 3.426%, 2.776%, and 21.576%. Besides, average alignment error of the left eye, right eye, nose, left mouth, and right mouth is performed on the proposed algorithm. (C)& nbsp;2022 Published by Elsevier B.V.
查看更多>>摘要：Electroencephalogram (EEG) is a noninvasive method to detect spatio-temporal electric signals in human brain, actively used in the recent development of Brain Computer Interfaces (BCI). EEG's patterns are affected by the task, but also other variable factors influence the subject focus on the task and result in noisy EEG signals difficult to decipher. To surpass these limitations methods based on artificial neural networks (ANNs) are used, they are inherently robust to noise and do not require models. However, they learn from examples and require lots of training data-sets. This will increase costs, need research time and subjects effort. To reduce the number of experiments necessary for network training, we devised a methodology to provide artificial data from a limited number of training data-sets. This was done by applying Empirical Mode Decomposition (EMD) on the EEG frames and intermixing their Intrinsic Mode Function (IMFs). We experimented on motor imagery (MI) tests where participants were asked to imagine movement of the left (or right) arm while under EEG recording. The EEG data were firstly transformed using the Morlet wavelet and then fed to an originally designed Convolutional Neural Network (CNN) with long short term memory blocks (LSTM-RNN). The introduction of artificial frames improved performances when compared with standard algorithms. The artificial frames become advantageous even when the number of available real frames was only of 7 or 8. In a test with two subjects (200 recordings for each subject), we reached an accuracy better than 88% for both subjects. Improvements due to the artificial data were especially noticeable for the under-performing subject, whose EEG had lower accuracy. Imagination recognition accuracy was about 89% with 360 training frames, in which 300 were artificially created starting from 60 real ones. We believe this methodology of synthesizing artificial data may contribute to the development of novel and more efficient ways to train neural networks for brain computer interfaces.(C) 2022 Elsevier B.V. All rights reserved.
查看更多>>摘要：The eruption of COVID-19 at the beginning of 2020 has sounded the alarm, making experts pay more attention to public health emergency events. A suitable emergency response plan plays a vital role in handling emergency events. Therefore, this paper focuses on the evaluation of emergency response plans among a set of group in the comprehensive prospect, and an emergency decision making method integrated with the interval type-2 fuzzy information based on the third generation prospect theory (PT3) and the extended MULTIMOORA method is proposed. Individuals express their preferences using some given linguistic terms set. Furthermore, considering the conflicts may occur in the group, a convergent iterative algorithm is designed for group consensus reaching. Then, the stochastic multi-criteria acceptability analysis (SMAA) method and the Borda Count (BC) method are generated to combine the results instead of the dominance theory in MULTIMOORA system. Finally, based on the background of the COVID-19 pandemic from Wuhan, a case study about the selection of emergency response plan and the corresponding sensitivity and comparative analysis are exhibited to explain the effectiveness of the proposed method. (C)& nbsp;2022 Elsevier B.V. All rights reserved.
查看更多>>摘要：Wind energy is a core sustainable source of electric power, and accurate wind-speed forecasting is pivotal to enhancing the power stability, efficiency, and utilization. The existing forecasting methods are still limited by the influence of outliers and the modelling difficulties caused by complex features in wind speed series. This paper proposes a new wind speed forecasting system based on a designed adaptive robust extreme learning machine (ARELM) model and signal decomposition algorithms. Firstly, the ARELM is designed to sufficiently lessen the violation of normality assumptions and contamination by outliers. ARELM takes an adaptive scaled Huber's loss as its objective function, which can limit the influence of outliers and adaptively determine an appropriate mixture distribution of normal distribution and Laplace distribution at the same time. Secondly, the empirical mode decomposition (EMD) method and its improved methods (EEMD, CEEMD and CEEMDAN) are introduced to our windspeed forecasting system, where the low-frequent sub-series are modelled by basic ELM and the high-frequent ones are modelled by ARELM. This can decompose the modelling complex wind speed series into modelling several simple sub-series and reduce the difficulty of modelling. Experimental results show that our combined forecasting system, ELM-ARELM, obtains up to 78% improvement in forecasting performance comparing with the methods using general Huber's loss and other comparison methods, which show the superiority of the adaptive scaled Huber's loss. The error indexes (MAE and RMSE) by the proposed system, which are (0.25, 0.34), (0.32, 0.45) and (0.38, 0.53) for 5 min head, 15 min ahead and 25 min ahead experiments respectively, demonstrate the effectiveness of decomposition methods on improving accuracy of wind speed prediction. (c) 2022 Elsevier B.V. All rights reserved.
查看更多>>摘要：Human Learning Optimization (HLO) is a simple yet powerful meta-heuristic developed based on a simplified human learning model. Many cognitive activities of humans contain an element of reasoning, and with reasoning, humans can gain deeper information on problems to boost learning performance. Inspired by this fact, this paper proposes a novel human learning optimization algorithm with reasoning learning (HLORL), in which a social reasoning learning operator (SRLO) is developed by using multiple social information sources to improve the global search ability of the algorithm. A parameter study is performed to give the recommended values of the control parameters. It also analyzes and discusses the role and function of the social reasoning learning operator. Finally, the proposed HLORL is applied to solve the CEC14 benchmark functions and 0-1 knapsack problems. The performance of HLORL is compared with the previous HLO variants and other state-of-art metaheuristics. The experimental results demonstrate that the proposed HLORL has significant advantages over the compared algorithms.
查看更多>>摘要：Solving nonlinear equations (NEs) has been obtained considerable attentions in recent years. However, it is still a difficult problem to improve the efficiency of the algorithm to find multiple roots of NEs. Aiming to deal with this issue, an archive guided speciation-based differential evolution (AGSDE) is presented in this paper. It contains three main components: (i) an archive construction approach is used to save the historical individual with poor fitness values in the selection phase; (ii) a reusing historical individual mechanism is implemented to guide the evolution; (iii) a local search method for solving NEs is performed on different subpopulations to refine the accuracy of the candidate solutions. The performance of AGSDE is tested on 30 NEs problems with different characteristics. Experimental results of AGSDE are competitive with those of other state-of-the-art methods in terms of root rate and success rate. In addition, AGSDE also shows its superiority for solving the other 10 complex NEs problems.(c) 2022 Published by Elsevier B.V.
查看更多>>摘要：Big data analytics in healthcare is emerging as a promising field to extract valuable information from large databases and enhance results with fewer costs. Although numerous methods have been proposed for big data analytics in the medical field, an authorized entity is required to access data, inhibiting diagnosis accuracy and efficiency. Particularly, the detection of lung cancer is critical as it is the third most common type of cancer occurring in both males and females in the US and a leading cause of cancer-related deaths worldwide, the detection of lung cancer. Therefore, this study introduces the Multivariate Ruzicka Regressed eXtreme Gradient Boosting Data Classification (MRRXGBDC) technique and service-oriented architecture (SOA) to improve the prediction accuracy and reduce the prediction time of lung cancer in big data analytics. Service-oriented architectures (SOAs) provide a set of healthcare services, where patient data are stored in the database of a physician or other certified entity. After receiving the patient data as input, several multivariate Ruzicka logistic regression trees are constructed by the physician to calculate the relationship between the dependent and independent variables. With this regression analysis, the presence or absence of disease is discovered. The experimental results reveal that the MRRXGBDC technique performs better with 10% improvement in prediction accuracy, 50% reduction of false positives, and 11% faster prediction time for lung cancer detection compared to existing works. (C)& nbsp;2022 Published by Elsevier B.V.
查看更多>>摘要：Earthquake prediction (EQP) is crucial for taking preemptive measures and accurately predicting damage. Several historical seismic-event-based EQP approaches have been proposed; however, these approaches only identify anomalies without distinguishing noise, reducing the prediction accuracy. Macrophages play an important role in the immune system by recognizing viruses, apoptotic cells, and normal cells, as well as performing immune responses and suppression to ensure homeostasis; that is, macrophages exhibit strong classification capabilities and self-adaptability. Therefore, in this study, a novel artificial macrophage algorithm (AMA) for EQP is proposed. More specifically, we first introduce the biological mechanism of macrophages to establish recognition and learning mechanisms to identify noise and anomalies. Second, we adopt a distance metric to denote the weights of the AMA, instead of using experience-based parameters. Finally, a stochastic gradient descent is introduced to ensure the adaptability of the AMA. The performance of the AMA was assessed through an analysis of historical seismic events in Sichuan and its surroundings. Our experimental results demonstrate that AMA outperforms state-of-the-art EQP algorithms. The parameters and statistical tests of AMA were further analyzed in this study. (C)& nbsp;2022 Elsevier B.V. All rights reserved.
NSTL
Elsevier