Article
Computer Science, Artificial Intelligence
Sarah Almaghrabi, Mashud Rana, Margaret Hamilton, Mohammad Saiedur Rahaman
Summary: Accurate and reliable prediction of photovoltaic power output is crucial for grid stability and power dispatching. Current wavelet transform methods have limitations in terms of time complexity. This study proposes a new approach that improves the efficiency of wavelet transform by using a simplified model.
Article
Computer Science, Artificial Intelligence
Hossein Abbasian Mohammadi, Sedigheh Ghofrani, Ali Nikseresht
Summary: Many studies on time series forecasting have used fuzzy cognitive maps (FCMs), but there is a need for techniques that can effectively respond to and accurately predict large-scale non-stationary time series, such as sharp-fluctuated datasets. This study proposes a combined forecasting framework with ridge regression, high-order FCM (HFCM), and empirical wavelet transform (EWT) to address this issue. The proposed method, named EWTHFCM, transforms non-stationary time series using EWT and models each multivariate time series through a node in HFCM. Ridge regression is employed for optimized learning and the inverse EWT is used to reconstruct the multivariate time series for prediction. The experimental results and comparison with other research papers suggest that EWTHFCM is a new, superior, and accurate method for various forecasting purposes and circumstances.
APPLIED SOFT COMPUTING
(2023)
Article
Computer Science, Artificial Intelligence
Aasim, S. N. Singh, Abheejeet Mohapatra
Summary: This paper proposes a hybrid model that combines the Wavelet Transform (WT) and Support Vector Machine (SVM) features for electrical load forecasting. The model improves overall forecasting ability by maximizing error contribution in sub-series forecasting.
APPLIED SOFT COMPUTING
(2021)
Article
Business, Finance
Amine Mtiraoui, Heni Boubaker, Lotfi BelKacem
Summary: This study proposes an innovative hybrid model, combining autoregressive fractionally integrated moving average (ARFIMA), empirical wavelet (EW) transform, and local linear wavelet neural network (LLWNN) approaches, to predict Bitcoin returns and volatilities. The model integrates the advantages of the long memory model, EW decomposition technique, artificial neural network structure, and back propagation and particle swarm optimization learning algorithms. Experimental results show that the optimized hybrid approach outperforms classic models in terms of providing accurate out-of-sample forecasts over longer horizons. The model proves to be the most appropriate Bitcoin forecasting technique, with smaller prediction errors compared to other computing techniques.
RESEARCH IN INTERNATIONAL BUSINESS AND FINANCE
(2023)
Article
Mathematics
Yuxin Zhang, Yifei Yang, Xiaosi Li, Zijing Yuan, Yuki Todo, Haichuan Yang
Summary: The famous McCulloch-Pitts neuron model is considered too simplistic in the long term. In contrast, the dendritic neuron model (DNM) is effective in prediction problems and captures the nonlinear information processing of synapses and dendrites. However, finding an efficient learning approach for DNMs remains challenging due to problems with the classical error back-propagation (BP) algorithm. This study classifies meta-heuristic algorithms (MHAs) into different clusters with different population interaction networks (PINs) and tests the performance of DNMs optimized by these clusters in financial time-series forecasting. The DNM optimized by MHAs with power-law-distributed PINs outperforms the DNM trained using the BP algorithm.
Article
Computer Science, Artificial Intelligence
Lifan Long, Qian Liu, Hong Peng, Qian Yang, Xiaohui Luo, Jun Wang, Xiaoxiao Song
Summary: A novel time series forecasting approach based on nonlinear spiking neural P systems is proposed in this study. By converting the time series into the frequency domain and automatically constructing and training NSNP systems in the frequency domain, sequence data for future time can be predicted. Experimental results demonstrate the availability and effectiveness of this approach on multiple time series datasets.
INTERNATIONAL JOURNAL OF NEURAL SYSTEMS
(2022)
Article
Engineering, Multidisciplinary
W. A. Shaikh, S. F. Shah, S. M. Pandhiani, M. A. Solangi
Summary: This investigative study focuses on the impact of wavelet on traditional forecasting time-series models, and finds that combining wavelet algorithms with traditional models can improve the accuracy of the forecasts.
CMES-COMPUTER MODELING IN ENGINEERING & SCIENCES
(2022)
Article
Engineering, Biomedical
C. Rahul, T. Arathi, Lakshmi S. Panicker, R. Gopikakumari
Summary: Machine translation is the process of converting one language to another using computers. Neural machine translation has significantly improved translation quality, but struggles with the complexity of Indian languages. This study proposes combining neural machine translation with morphology, part of speech tagging, and word sense disambiguation for bidirectional translation between Sanskrit and Malayalam. Experiments show that adding a speech modality improves the overall quality of translation, with BLEU scores of 43.89 for Sanskrit to Malayalam and 42.72 for Malayalam to Sanskrit multimodal translation.
BIOMEDICAL SIGNAL PROCESSING AND CONTROL
(2023)
Article
Engineering, Civil
Mostafa Rezaali, John Quilty, Abdolreza Karimi
Summary: The study aims to forecast short-term urban water demand using the Wavelet Data-Driven Forecasting Framework and examines the effectiveness of different machine learning models, dataset partitioning methods, and input variable selection approaches. Real-world case studies show that probabilistic RF and its 'best' wavelet-based version provide the most accurate and reliable forecasts, while permutation- and bootstrap-based dataset partitioning approaches have potential to reduce overfitting. Wavelet decomposition improves model performance and RFIVS significantly reduces the number of input variables used in the ML models while improving performance.
JOURNAL OF HYDROLOGY
(2021)
Article
Ecology
Celso Augusto Guimaraes Santos, Gleycielle Rodrigues do Nascimento, Camilo Allyson Simoes de Farias, Richarde Marques da Silva, Manoranjan Mishra
Summary: This study presents a methodology that improves streamflow forecasting by using wavelet neural networks to relate streamflows with rainfall data. The methodology performs well in long-term forecasts, especially in the Mahanadi River basin, and can be applied in other catchments.
ECOLOGICAL INFORMATICS
(2023)
Article
Computer Science, Artificial Intelligence
Donghwan Kim, Jun-Geol Baek
Summary: This paper proposes a method to improve the performance of predictive models in time series data. The method decomposes the data using the maximum overlap discrete wavelet transform (MODWT) and uses bootstrap to generate bootstrapped data. Experimental results show that the proposed method can improve the performance of existing algorithms, especially when the algorithm is simple.
EXPERT SYSTEMS WITH APPLICATIONS
(2022)
Article
Mathematics
Andres M. Kowalski, Mariela Portesi, Victoria Vampa, Marcelo Losada, Federico Holik
Summary: This study investigates the information contained in COVID-19 data of infected and deceased individuals across all countries using informational quantifiers such as entropy and statistical complexity. The results provide insights into the available information and can contribute to addressing this global issue.
Article
Computer Science, Information Systems
Mario Maya, Wen Yu
Summary: This paper proposes a hybrid Meta-Transfer Learning technique based on transfer-learning, meta-learning and signal detection to solve the problems of data loss in data acquisition and long-term forecast in multi-horizon time series forecasting. The effectiveness of the method is validated by predicting earthquakes magnitude in Italy.
Article
Computer Science, Artificial Intelligence
Ruobin Gao, Liang Du, Kum Fai Yuen, Ponnuthurai Nagaratnam Suganthan
Summary: The study proposes a hybrid model that combines Empirical Wavelet Transformation (EWT) with Random Vector Functional Link (RVFL) to enhance multi-scale feature extraction ability. The empirical study demonstrates that the hybrid model achieves high accuracy and avoids data leakage issues during forecasting.
APPLIED SOFT COMPUTING
(2021)
Article
Engineering, Civil
Rahim Barzegar, Mohammad Taghi Aalami, Jan Adamowski
Summary: This study demonstrated the importance of accurate lake water level forecasting models for various applications. By combining BC-MODWT data preprocessing with a hybrid CNN-LSTM deep learning model, improved accuracy in Lake Michigan and Lake Ontario water level predictions was achieved, surpassing traditional machine learning models. The proposed BC-MODWT-CNN-LSTM model proved to be a potentially useful approach for enhancing the accuracy of lake water level forecasts.
JOURNAL OF HYDROLOGY
(2021)
Article
Green & Sustainable Science & Technology
Mohammad Ehteram, Ali Najah Ahmed, Chow Ming Fai, Sarmad Dashti Latif, Kwok-wing Chau, Kai Lun Chong, Ahmed El-Shafie
Summary: This research aims to optimize water release for optimal hydropower generation in the future. The results show that temperature will increase while precipitation will decrease. The adoption of adaptive rule curves can improve hydropower generation and reduce vulnerability in the future period.
ENVIRONMENT DEVELOPMENT AND SUSTAINABILITY
(2023)
Article
Meteorology & Atmospheric Sciences
Francis Polong, Khidir Deng, Quoc Bao Pham, Nguyen Thi Thuy Linh, S. I. Abba, Ali Najah Ahmed, Duong Tran Anh, Khaled Mohamed Khedher, Ahmed El-Shafie
Summary: This study evaluates the effects of land use/land cover (LULC) and climate changes on hydrological processes in a tropical catchment. The Soil and Water Assessment Tool (SWAT) model is used to simulate different combinations of LULC and climate scenarios. The results show that climate changes have a greater impact on the simulated parameters than changes in LULC. The study emphasizes the need to assess the isolated and combined effects of LULC and climatic changes when evaluating impacts on hydrological processes.
THEORETICAL AND APPLIED CLIMATOLOGY
(2023)
Article
Water Resources
Afif Fitri Aziz, Nurul Hani Mardi, Marlinda Abdul Malek, Su Yean Teh, Mohd Azwan Wil, Abd Halim Shuja, Ali Najah Ahmed, Pavitra Kumar, Mohsen Sherif, Ahmed Elshafie
Summary: The coastal zone is economically important and experiences increasing development and activities. Tsunamis, often caused by oceanic earthquakes, pose a significant threat to coastal areas, as demonstrated by the 2004 Andaman tsunami in Malaysia. This study investigates the potential seismic activities and simulates Manila Trench earthquake-induced tsunamis of different intensities on the East Coast of Peninsular Malaysia. Findings reveal that a Mw 9.0 earthquake results in the most disastrous effects, with Kelantan experiencing the highest inundation depth of 4.0 m in Pasir Puteh and Terengganu experiencing the highest inundation depth of 6.0 m in Kuala Terengganu. Mitigation measures and evacuation plans can be improved based on these findings to minimize property and life losses.
APPLIED WATER SCIENCE
(2023)
Article
Water Resources
Wei Joe Wee, Kai Lun Chong, Ali Najah Ahmed, Marlinda Binti Abdul Malek, Yuk Feng Huang, Mohsen Sherif, Ahmed Elshafie
Summary: Hydrologists rely heavily on river streamflow prediction for flood management and water demand monitoring. In this study, a hybrid model combining bat algorithm and artificial neural network was used to optimize streamflow forecasting, showing superior performance compared to traditional artificial neural network models.
APPLIED WATER SCIENCE
(2023)
Article
Water Resources
Wan Norsyuhada Che Wan Zanial, Marlinda Binti Abdul Malek, Mohd Nadzri Md Reba, Nuratiah Zaini, Ali Najah Ahmed, Mohsen Sherif, Ahmed Elshafie
Summary: In this research, the integration of Artificial Neural Network (ANN) with Cuckoo search algorithm (CS-ANN) was used to accurately estimate the effect of changing rainfall patterns on river flow. The proposed Hybrid CS-ANN model showed improved accuracy in predicting river flow compared to the stand-alone ANN model.
APPLIED WATER SCIENCE
(2023)
Article
Water Resources
K. L. Chong, Y. F. Huang, C. H. Koo, Mohsen Sherif, Ali Najah Ahmed, Ahmed El-Shafie
Summary: Streamflow forecasting is crucial in water resources management, and this paper explores the use of machine learning algorithms for two distinct streamflow forecasting problems. The study finds that categorical-based streamflow forecast outperforms regression-based forecast, and forest-based algorithms are superior for predicting high streamflow fluctuations with low-dimensional input. Furthermore, encoding streamflow time series as images for forecasting demands further analysis as different approaches yield varying results.
APPLIED WATER SCIENCE
(2023)
Review
Green & Sustainable Science & Technology
Sarmad Dashti Latif, Ali Najah Ahmed
Summary: This review paper explores the use of deep learning and machine learning algorithms for reservoir inflow prediction in hydrological forecasting. It analyzes the application of AI models in various hydrology sectors and focuses on the two primary categories of deep learning and machine learning. The study examines the long short-term memory deep learning method and three traditional machine learning algorithms, and provides a summary of the findings, benefits, and drawbacks discovered through literature reviews.
ENVIRONMENT DEVELOPMENT AND SUSTAINABILITY
(2023)
Article
Water Resources
Jing Lin Ng, Yuk Feng Huang, Aik Hang Chong, Jin Chai Lee, Muyideen Abdulkareem, Nur Ilya Farhana Md Noh, Majid Mirzaei, Ali Najah Ahmed
Summary: This study evaluated the main environmental issue of drought in Peninsular Malaysia and conducted a temporal analysis of multiple drought indices. The findings suggest that shorter timescales are more suitable for assessing short-term droughts. The Z-score index was found to have the highest accuracy among all the indices, which is valuable in accurately evaluating drought indices.
JOURNAL OF WATER AND CLIMATE CHANGE
(2023)
Article
Multidisciplinary Sciences
Ahmad Danboos, Suraya Sharil, Firdaus Mohamad Hamzah, Ayman Yafouz, Yuk Feng Huang, Ali Najah Ahmed, Abdel Azim Ebraheem, Mohsen Sherif, Ahmed El-Shafie
Summary: Iraq is facing a severe water crisis as the flow of water in the Tigris and Euphrates Rivers decreases. Due to population growth, studies estimate a water shortage of 44 billion cubic meters by 2035. In response, the Water Budget-Salt Balance Model (WBSBM) has been developed to calculate the potential net water saving from Non-Conventional Water Resources (NCWRs). The WBSBM consists of four stages, including identifying data, demonstrating water user activities, developing the model, and computing net water saving. The results show potential net water savings of 6.823 and 6.626 billion cubic meters per year in 2025 and 2035, respectively. The WBSBM model provides comprehensive examination and determination of optimal net water saving amounts through various scenarios of utilizing NCWRs.
Article
Engineering, Multidisciplinary
Nur Nazmi Liyana Mohd Napi, Samsuri Abdullah, Amalina Abu Mansor, Nurul Adyani Ghazali, Ali Najah Ahmed, Nazri Che Dom, Marzuki Ismail
Summary: Based on analysis of five years of meteorological and gaseous pollutants data, this study developed three different O3 concentration prediction models, with MLR2 model being considered the best due to its lowest root mean square error and mean absolute error. The establishment of an O3 prediction model can provide local governments with early information to help them reduce and manage air pollution emissions.
INTERNATIONAL JOURNAL OF INTEGRATED ENGINEERING
(2023)
Article
Engineering, Multidisciplinary
Marwah Sattar Hanoon, Ali Najah Ahmed, Arif Razzaq, Atheer Y. Oudah, Ahmed Alkhayyat, Yuk Feng Huang, Pavitra Kumar, Ahmed El-Shafie
Summary: This study investigates the capability of various machine learning algorithms in predicting the power production of a reservoir located in China. The proposed models can efficiently predict the hydropower generation and provide valuable insights for energy decision-makers.
AIN SHAMS ENGINEERING JOURNAL
(2023)