Article
Computer Science, Artificial Intelligence
Rakshitha Godahewa, Kasun Bandara, Geoffrey Webb, Slawek Smyl, Christoph Bergmeir
Summary: Ensembling techniques are used to improve the performance of Global Forecasting Models (GFM) and univariate models in heterogeneous datasets. A new clustered ensembles methodology is proposed to train multiple GFMs on different clusters of series, achieving higher accuracy than baseline models.
KNOWLEDGE-BASED SYSTEMS
(2021)
Article
Computer Science, Artificial Intelligence
Yong Li, Richard Gault, T. Martin McGinnity
Summary: The article introduces a probabilistic fuzzy neural algorithm with a recurrent probabilistic generation module (PFNN-R) to enhance the ability of PFNNs to accommodate noisy data. The back-propagation-based mechanism is utilized to shape the distribution of the probabilistic density function of the fuzzy membership. Through simulation results, it is demonstrated that incorporating recurrency advances the ability of PFNNs to model time-series data with high intensity, random noise.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2022)
Article
Multidisciplinary Sciences
Zhaohui Xia, Lei Qin, Zhen Ning, Xingyu Zhang
Summary: This paper presents different prediction models using deep learning methods based on the monthly incidence of Hepatitis in China. The performance of three deep learning methods, LSTM, RNN, and BPNN, are assessed and compared. The results show that no single model is superior in predicting the incidence cases of Hepatitis. These deep learning time series predictive models are significant for forecasting the Hepatitis incidence and can assist decision-makers in making efficient decisions for early detection and control of the disease.
Article
Computer Science, Information Systems
Rohitash Chandra, Shaurya Goyal, Rishabh Gupta
Summary: This study evaluates the performance of deep learning models for multi-step ahead time series prediction, and finds that bidirectional and encoder-decoder LSTM networks show the best accuracy in given time series problems.
Article
Computer Science, Artificial Intelligence
Jingyuan Wang, Zhen Peng, Xiaoda Wang, Chao Li, Junjie Wu
Summary: The article introduces a novel extension of Fuzzy Cognitive Map called Deep FCM for multivariate time series forecasting, combining the predictive advantage of deep neural networks and the interpretative advantage of FCM. DFCM utilizes fully connected neural networks and recurrent neural networks to model concept relationships and external factors in the system, and proposes a partial derivative-based method to improve model interpretability.
IEEE TRANSACTIONS ON FUZZY SYSTEMS
(2021)
Article
Engineering, Environmental
Malte Jahn, Christian H. Weiss
Summary: Despite being relevant in various areas, there are few stochastic models for ordinal time series discussed in the literature. This study proposes different ordinal GARCH-type models to accommodate flexible serial dependence structure and handle nonlinear dependence and intensified memory. The logistic ordinal GARCH model considers the natural order by relying on conditional cumulative distributions, while the conditionally multinomial model utilizes softmax response function to incorporate ordinal information by considering past categories. The study shows that the resulting neural softmax GARCH model, which combines the latter model with artificial neural network response function, offers great flexibility and brings benefits in real-world applications.
STOCHASTIC ENVIRONMENTAL RESEARCH AND RISK ASSESSMENT
(2023)
Article
Computer Science, Information Systems
Francisco Martinez, Maria P. Frias, Maria D. Perez-Godoy, Antonio J. Rivera
Summary: This paper explores the use of a pool of time series to predict individual series and proposes several approaches, including training models with mutually exclusive series and combining their forecasts. The experimental results using generalized regression neural networks are promising, and the approaches allow for forecasting series that are too short for traditional models.
Article
Computer Science, Information Systems
Michael Gallagher, Nikolaos Pitropakis, Christos Chrysoulas, Pavlos Papadopoulos, Alexios Mylonas, Sokratis Katsikas
Summary: Machine learning and Artificial Intelligence (AI) are already assisting human decision-making and have the potential to make autonomous decisions in the future. However, adversarial machine learning attacks can significantly impact the accuracy of these systems due to their sensitivity to modified input data.
COMPUTERS & SECURITY
(2022)
Article
Automation & Control Systems
Martha Ramirez, Patricia Melin
Summary: This research combines neural networks with type-2 fuzzy systems to perform clustering and prediction of time series data related to population, urban population, PM2.5, CO2, registered cases, and deaths from COVID-19 for certain countries. The approach simulates the behavior of cognitive functions in the human brain using different types of neural models and interval type-2 fuzzy logic for decision-making process. The results show the advantages of this approach in clustering, predicting, and integrating multiple time series data for decision-making with uncertainty.
INTERNATIONAL JOURNAL OF FUZZY SYSTEMS
(2023)
Article
Multidisciplinary Sciences
Vitor Hugo Serravalle Reis Rodrigues, Paulo Roberto de Melo Barros Junior, Euler Bentes dos Santos Marinho, Jose Luis Lima de Jesus Silva
Summary: Developing accurate models for groundwater control is crucial for managing and planning water resources from aquifer reservoirs. The proposed Wavelet Gated Multiformer combines the strengths of a vanilla Transformer and a Wavelet Crossformer to improve the model's predictive capabilities by computing the relationships between time-series points and finding trending periodic patterns. This model outperforms other transformer-like models in terms of Mean Absolute Error reduction.
SCIENTIFIC REPORTS
(2023)
Article
Mathematics, Applied
Xinhui Li
Summary: This paper proposes a financial forecasting method that combines support vector machine with convolutional neural network model and applies it to predict the trend of stock indices. The experimental results demonstrate that the proposed model can more accurately predict the trend of stock indices.
JOURNAL OF FUNCTION SPACES
(2022)
Article
Multidisciplinary Sciences
Alvaro David Orjuela-Canon, Andres Leonardo Jutinico, Mario Enrique Duarte Gonzalez, Carlos Enrique Awad Garcia, Erika Vergara, Maria Angelica Palencia
Summary: This study used artificial neural networks to predict the development trend of tuberculosis and found that traditional models performed better, which can help health authorities propose more effective control strategies.
Article
Computer Science, Artificial Intelligence
Kasun Bandara, Christoph Bergmeir, Hansika Hewamalage
Summary: The article introduces a unified prediction framework (LSTM-MSNet) based on decomposition for forecasting time series with multiple seasonal patterns. Unlike traditional methods, this framework globally trains a model to utilize knowledge from all related time series. Experiments show that decomposition is beneficial for datasets from different sources, while exogenous seasonal variables or no seasonal preprocessing may be better choices for homogeneous series in practical applications.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2021)
Article
Engineering, Civil
Julien Monteil, Anton Dekusar, Claudio Gambella, Yassine Lassoued, Martin Mevissen
Summary: This work investigates the use of deep learning models for long-term large-scale traffic prediction tasks, focusing on scalability. By analyzing 14 weeks of speed observations from over 1000 segments in downtown Los Angeles, different machine learning and deep learning predictors were studied, along with their scalability to larger areas. The study shows that modeling temporal and spatial features into deep learning predictors can be beneficial for long-term predictions, while simpler predictors achieve satisfactory performance for link-based and short-term forecasting, with a trade-off in prediction accuracy, horizon, training time, and model sizing discussed.
IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS
(2022)
Article
Automation & Control Systems
Yin Sheng, Tingwen Huang, Zhigang Zeng
Summary: This article investigates global exponential stabilization of Takagi-Sugeno fuzzy memristive neural networks with multiple time-varying delays via intermittent control strategy. The outcome is generalized to networks with infinite distributed time delays, and the global exponential stability of networks with discrete time-varying delays is explored in terms of 1-norm. The derived conditions contain certain existing ones as special cases, and examples are presented to validate the results.
IEEE TRANSACTIONS ON SYSTEMS MAN CYBERNETICS-SYSTEMS
(2022)
Article
Computer Science, Artificial Intelligence
Debabrata Dansana, Raghvendra Kumar, Aishik Bhattacharjee, D. Jude Hemanth, Deepak Gupta, Ashish Khanna, Oscar Castillo
Summary: The COVID-19 pandemic, which began in December 2019 in China, has rapidly spread worldwide and infected over ten million people. This study explores the binary classification of pneumonia using convolution neural networks on X-ray and CT scan images. The findings show that fine-tuned versions of VGG-19 and Inception_V2 models demonstrate high accuracy rates.
Article
Computer Science, Artificial Intelligence
Patricia Melin, Julio Cesar Monica, Daniela Sanchez, Oscar Castillo
Summary: In this paper, we propose a hybrid ensemble modular neural network approach for predicting the COVID-19 time series worldwide. This approach combines nonlinear autoregressive neural networks to form efficient predictors for each country. The analysis using publicly available datasets reveals interesting conclusions that can aid countries in devising effective strategies for combating the pandemic. The proposed approach may also offer guidance for similar countries.
Article
Computer Science, Artificial Intelligence
Patricia Melin, Daniela Sanchez, Julio Cesar Monica, Oscar Castillo
Summary: This paper presents a method for predicting global COVID-19 pandemic using a firefly algorithm to design an ensemble neural network architecture. It takes into account the differences between countries and utilizes type-2 fuzzy logic and weighted average integration to improve prediction accuracy.
Article
Automation & Control Systems
Martha Ramirez, Patricia Melin
Summary: This research combines neural networks with type-2 fuzzy systems to perform clustering and prediction of time series data related to population, urban population, PM2.5, CO2, registered cases, and deaths from COVID-19 for certain countries. The approach simulates the behavior of cognitive functions in the human brain using different types of neural models and interval type-2 fuzzy logic for decision-making process. The results show the advantages of this approach in clustering, predicting, and integrating multiple time series data for decision-making with uncertainty.
INTERNATIONAL JOURNAL OF FUZZY SYSTEMS
(2023)
Article
Chemistry, Analytical
Hector Carreon-Ortiz, Fevrier Valdez, Patricia Melin, Oscar Castillo
Summary: Recurrent Neural Networks (RNN) are commonly used for time series and sequential data applications and are currently used in embedded devices. However, RNNs have the drawbacks of high computational cost and memory requirements. This article explores the use of Nonlinear Autoregressive Neural Networks (NARNN), a type of RNN, and applies the Discrete Mycorrhizal Optimization Algorithm (DMOA) to optimize the NARNN architecture. The proposed approach achieves good results when tested with the Mackey-Glass chaotic time series (MG), and comparisons with other methods like Backpropagation and ANFIS also yield positive outcomes. This algorithm has potential applications in various fields including robotics, microsystems, sensors, and 3D printing.
Article
Automation & Control Systems
Patricia Ochoa, Oscar Castillo, Patricia Melin, Juan R. Castro
Summary: This article presents the usage of interval type-3 fuzzy sets in the differential evolution algorithm for the first time. A study is conducted to explore the influence of the LowerScale (lambda) parameter on the convergence of differential evolution. The results from experiments on benchmark functions and motor control optimization demonstrate that the combination of interval type-3 fuzzy sets and differential evolution outperforms type-1 and interval type-2 variants.
INTERNATIONAL JOURNAL OF FUZZY SYSTEMS
(2023)
Article
Multidisciplinary Sciences
Mario Garcia-Valdez, Alejandra Mancilla, Oscar Castillo, Juan Julian Merelo-Guervos
Summary: In this work, a distributed and asynchronous bio-inspired algorithm is proposed to speed up the design process of a controller by executing simulations in parallel. The algorithm uses a multi-population multi-algorithmic approach with isolated populations interacting asynchronously using a distributed message queue. The results demonstrate the speedup benefit of the proposed algorithm and the advantages of mixing populations with distinct metaheuristics.
Editorial Material
Computer Science, Artificial Intelligence
Oscar Castillo, Oscar Montiel, Fevrier Valdez
Retraction
Computer Science, Artificial Intelligence
Patricia Melin, Julio Cesar Monica, Daniela Sanchez, Oscar Castillo
Article
Multidisciplinary Sciences
Lucio Amezquita, Oscar Castillo, Jose Soria, Prometeo Cortes-Antonio
Summary: This work introduces multiple variations of the Multi-verse Optimizer Algorithm (MVO) by incorporating chaotic maps to generate new solutions. The variations, called Fuzzy-Chaotic Multi-verse Optimizer (FCMVO), replace random values with chaotic maps from literature for certain parameters in the original algorithm. Fuzzy Logic is also used for dynamic parameter adaptation in these new variants, along with the analysis of the improvement over the Fuzzy MVO. The objective is to compare the performance of MVO algorithm with the best-performing chaotic maps and Fuzzy Logic in benchmark mathematical functions before exploring other case studies.
Article
Multidisciplinary Sciences
Oscar Castillo, Patricia Melin
Summary: This paper presents an initial proposal for the utilization of mediative fuzzy logic in control problems. It extends the concept of fuzzy control to mediative fuzzy logic for situations involving two or more control experts, aiming to improve control results by combining their knowledge. The study demonstrates the effectiveness of type-3 mediative fuzzy control in handling uncertainty from noise in the control process.
Article
Chemistry, Multidisciplinary
Cinthia Peraza, Patricia Ochoa, Oscar Castillo, Patricia Melin
Summary: This article proposes the use of the theory of shadowed type-2 fuzzy sets to address complex control problems and reduce computational costs. By employing two alpha planes in the harmony search algorithm, effective results can be obtained. The simulations demonstrate that including noise improves the system's performance.
APPLIED SCIENCES-BASEL
(2023)
Article
Computer Science, Artificial Intelligence
Oscar Castillo, Juan R. Castro, Patricia Melin
Summary: In this article, a design methodology for Mamdani interval and general type-2 fuzzy systems with center-of-sets type reduction is presented. The methodology utilizes descriptive statistics, fuzzy c means clustering, and granular computing theory to define the justifiable footprint of uncertainty (JFOU) of the fuzzy granules. The design methodology is presented in three general steps, focusing on building a diagram of the justifiable information granule, characterizing and parameterizing the asymmetric type-2 membership functions, and obtaining all the justifiable information fuzzy granules for the fuzzy model. Experiments were conducted to evaluate the reliability of the proposed methodology.
JOURNAL OF MULTIPLE-VALUED LOGIC AND SOFT COMPUTING
(2023)
Article
Engineering, Multidisciplinary
Deepshikha Sarma, Amrit Das, Oscar Castillo, Uttam Kumar Bera
Summary: This research proposes a mathematical model for the transportation of relief materials in disaster response, aiming to minimize total cost and maximize coverage of affected people. The model considers both certain and uncertain environments, and uses interval-coefficient decomposition, expected value operator, and rough chance constraint programming to handle uncertainty. A real-life problem of Assam flood is used to evaluate the efficiency of the model, and a comparative study is conducted. The model provides optimal results in crisp form in certain environment, and rough interval, crisp, and interval forms in uncertain environment for better decision-making.
SADHANA-ACADEMY PROCEEDINGS IN ENGINEERING SCIENCES
(2023)