Article
Environmental Sciences
Yuxing Li, Lili Liang, Shuai Zhang
Summary: The fractal dimension (FD) is a classical nonlinear dynamic index that reflects the dynamic transformation of a signal. However, it can only capture signal information of a single scale in the entire frequency band. To address this limitation, a refined composite multi-scale FD (RCMFD) is proposed, which combines refined composite multi-scale processing with FD to capture multi-scale signal information. Hierarchical RCMFD (HRCMFD) is introduced to represent multi-scale signal information in sub-frequency bands. Two ship-radiated noise (SRN) multi-feature extraction methods based on RCMFD and HRCMFD are proposed, which effectively discriminate different signals.
Article
Mathematics, Interdisciplinary Applications
Zbigniew Omiotek, Roza Dzierzak, Andrzej Kepa
Summary: Fractal analysis was utilized to extract feature descriptors for diagnosing bone damage caused by osteoporosis from CT images of vertebrae, resulting in three descriptors. The K-NN classifier demonstrated the highest overall classification accuracy among the supervised classification methods, showing promise for diagnosing osteoporotic bone defects.
FRACTALS-COMPLEX GEOMETRY PATTERNS AND SCALING IN NATURE AND SOCIETY
(2021)
Article
Computer Science, Artificial Intelligence
Harmandeep Kaur, Munish Kumar
Summary: The article introduces a holistic approach and XGBoost technique for recognizing offline handwritten Gurumukhi words, achieving high accuracy and performance by classifying the handwritten words based on extracted features.
Article
Agronomy
Qi Wang, Peng Guo, Shiwei Dong, Yu Liu, Yuchun Pan, Cunjun Li
Summary: Accurate extraction of cropland distribution information using remote sensing technology is achieved through a fractal algorithm that integrates temporal and spatial features. The method uses Landsat 8 OLI data to extract multi-seasonal fractal features, demonstrating its efficiency in black soil areas in Lishu County, Northeast China. The results show a high consistency with statistical data, indicating that this method can accurately monitor changes in cropland and support the conservation and development of black soil in China.
Article
Physics, Multidisciplinary
Afek Ilay Adler, Amichai Painsky
Summary: Gradient Boosting Machines (GBM) are popular algorithms for tabular data, but suffer from bias in their base learners. This paper investigates the effect of biased base learners on GBM feature importance measures and proposes a framework using unbiased base learners to address this issue at a low computational cost. The suggested framework significantly improves GBM feature importance measures without sacrificing prediction accuracy.
Article
Materials Science, Multidisciplinary
Hang Li, Ze Zhang, Jinbang Zhai, Linzhen Yang, Haichao Long
Summary: Soil adhesion refers to the ability of soil particles to adhere to an external medium after being in contact with water, and its strength is defined by the adhesive force, which is closely related to the material of the external medium and the soil type. This study measured the adhesive force using a small-scale model to investigate the effects of different particle sizes on soil adhesion. The correlation between soil adhesion and parameters was analyzed by calculating the fractal dimension. The results confirmed that the fractal dimension had an impact on the peak value of the adhesive force and moisture, and the soil clay content played a key role in soil adhesion.
Article
Environmental Sciences
Elise Van Eynde, Arthur Nicolaus Fendrich, Cristiano Ballabio, Panos Panagos
Summary: This study used a machine learning model to analyze the zinc concentration in 21,682 soil samples and created a map of topsoil zinc concentrations in Europe. The research found that clay content was the most important factor affecting the distribution of soil zinc in Europe, with lower concentrations in coarser soils. Additionally, deposits and mining activities were identified as the main reasons for relatively high zinc concentrations in certain areas. The map developed in this study is crucial for assessing eco-toxicological risks and zinc deficiency in European soils, as well as for future policy-making.
SCIENCE OF THE TOTAL ENVIRONMENT
(2023)
Article
Mathematics, Applied
Uta Freiberg, Stefan Kohl
Summary: The Lorenz attractor is a well known attractor of a dynamical system and its box dimension is investigated using a developed algorithmic approach. While the algorithm performs well for most examples such as self-similar fractals, it seems to fail for the Lorenz attractor. The notion of local box dimension is introduced to reveal a hidden multifractal structure of the Lorenz attractor.
COMMUNICATIONS IN NONLINEAR SCIENCE AND NUMERICAL SIMULATION
(2021)
Article
Geosciences, Multidisciplinary
Jianhua Ren, Ruifeng Xie, Honglei Zhu, Yue Zhao, Zhuopeng Zhang
Summary: The study analyzed the relationships between different types of crack parameters and the salinity of soil samples to establish optimal prediction models for electrical conductivity (EC) values. The results showed that the crack intensity factor (CIF) had a weaker correlation with EC values compared to fractal dimensions and GLCM texture features. Contrast (CON) texture feature was identified as the best predictor of EC with high accuracy.
Article
Soil Science
Yi Xiao, Jie Xue, Xianglin Zhang, Nan Wang, Yongsheng Hong, Yefeng Jiang, Yin Zhou, Hongfen Teng, Bifeng Hu, Emanuele Lugato, Anne C. Richer-de-Forges, Dominique Arrouays, Zhou Shi, Songchao Chen
Summary: This study evaluated the potential of machine learning methods in predicting mineral-associated organic carbon (MAOC) in soil. The results showed that machine learning-based predictive models can accurately predict MAOC, and the use of feature selection methods can optimize model performance and simplify model structure. Additionally, the study found that model ensemble methods can improve the accuracy and robustness of predictive models.
Article
Thermodynamics
Xiaosong Ding, Chong Feng, Peiling Yu, Kaiwen Li, Xi Chen
Summary: This paper investigates the real-time prediction of nitrogen oxides (NOX) emission using around 17000 samples from a waste incineration power plant. A hybrid procedure is developed to select appropriate features from the unsynchronized data and establish a model based on gradient boosting decision tree (GBDT). The computational experiments show that GBDT outperforms its popular counterparts, supporting vector regression (SVR) and long short-term memory (LSTM), with low root mean square error (RMSE) values for training and test data. Shapley additive explanations (SHAP) analysis is also conducted.
Article
Computer Science, Information Systems
Sheela Ramachandra, Suchithra Ramachandran
Summary: This paper proposes a Periocular recognition algorithm that utilizes region-specific and sub-image-based neighbor gradient feature extraction to achieve better recognition results. The proposed method segments the periocular region into sub-regions and extracts features using different algorithms. Experimental results demonstrate that the proposed method outperforms traditional algorithms on multiple datasets.
JOURNAL OF KING SAUD UNIVERSITY-COMPUTER AND INFORMATION SCIENCES
(2022)
Article
Computer Science, Artificial Intelligence
Resmiye Nasiboglu, Efendi Nasibov
Summary: This study explores the application of gradient boosting regression models in the presence of fuzzy target variables and compares different defuzzification methods. The results show that the WABL method, when properly adjusted, outperforms other defuzzification methods across all datasets.
EXPERT SYSTEMS WITH APPLICATIONS
(2023)
Article
Computer Science, Artificial Intelligence
Wanan Liu, Hong Fan, Meng Xia
Summary: Credit scoring is crucial for banks and lending companies to manage credit risk exposure and drive profits, with GBDTs showing promising improvement in this area. However, the diversity of individual classifiers in GBDT is limited by iterative modifications to fitting targets and working on the same features. The proposed AugBoost-RFS/AugBoost-RFU outperforms GBDT on large-scale credit scoring datasets, offering strong interpretability.
EXPERT SYSTEMS WITH APPLICATIONS
(2022)
Article
Computer Science, Artificial Intelligence
Sangdi Lin, Bahareh Azarnoush, George Runger
Summary: This paper proposes a multi-target boosting method, named MTBR, for regression problems. Although it builds models separately for each target attribute, all target attributes are utilized when building each model by selecting the best models from all target attributes in each boosting iteration. The novel knowledge transfer approach introduced in this method uses the tree structure learned from one target attribute to predict another, proving the effectiveness of MTBR in leveraging knowledge from multiple target attributes and improving model accuracy through experiments with six datasets.
IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING
(2021)
Article
Environmental Sciences
Min Ji, Lanfa Liu, Runlin Du, Manfred F. Buchroithner
Article
Environmental Sciences
Sekela Twisa, Manfred F. Buchroithner
Article
Environmental Sciences
Zeinab Shirvani, Omid Abdi, Manfred Buchroithner
Article
Environmental Sciences
Zeinab Shirvani, Omid Abdi, Manfred F. Buchroithner
LAND DEGRADATION & DEVELOPMENT
(2020)
Article
Chemistry, Multidisciplinary
Min Ji, Lanfa Liu, Rongchun Zhang, Manfred F. Buchroithner
APPLIED SCIENCES-BASEL
(2020)
Article
Environmental Sciences
Emil Bayramov, Manfred Buchroithner, Martin Kada
Article
Environmental Sciences
Emil Bayramov, Manfred Buchroithner, Martin Kada, Yermukhan Zhuniskenov
Summary: This research focused on quantitatively assessing surface deformation velocities and rates at Tengiz Oilfield in Kazakhstan using remote sensing techniques. The study revealed continuous subsidence at the oilfield due to natural and man-made factors, with maximum annual vertical deformation velocity reaching 70 mm. The results will support operators of oil and gas fields and other infrastructure in evaluating ground deformation measurements.
Article
International Relations
Ellison Shumba
Summary: This study explores the application of 'Wolf Warrior' diplomacy by Chinese diplomats in Africa on social media platforms like Twitter and Facebook, particularly focusing on their posts regarding the US government and the pandemic in early 2020. By analyzing the social media accounts of former Chinese Ambassador to South Africa Lin Songtian and others, the study aims to understand the implications of this diplomatic approach on Africa and the emerging practice of African digital diplomacy.
SOUTH AFRICAN JOURNAL OF INTERNATIONAL AFFAIRS-SAJIA
(2021)
Article
Materials Science, Paper & Wood
Mohammad Farajollah Pour, Meysam Mehdinia, Mohammad Valizadeh Kiamahalleh, Kazem Doost Hoseini, Hamid Hatefnia, Ali Dorieh
Summary: The study investigated the biological durability of particleboard containing silver and copper nanoparticles when exposed to the white-rot fungus Trametes versicolor. Results showed that the samples with 15% silver nanoparticles exhibited the highest resistance to fungal attack. The presence of nanoparticles on the wood particles' surface acted as an impressive barrier to inhibit fungal growth, while also improving the internal bonding properties of the particleboard.
WOOD MATERIAL SCIENCE & ENGINEERING
(2022)
Article
Veterinary Sciences
David De la Torre, Claudete S. Astolfi-Ferreira, Ruy D. Chacon, Antonio J. Piantino Ferreira
Summary: Reports of chicken astrovirus (CAstV) infections have increased in recent years, affecting the poultry industries in Brazil and worldwide. This study characterized the CAstV strains found in breeding and incubation companies and revealed a potential association with white chick syndrome.
Article
Environmental Sciences
Emil Bayramov, Giulia Tessari, Martin Kada, Saida Aliyeva, Manfred Buchroithner
Summary: This study assessed differential vertical and horizontal deformations for the offshore Kashagan oilfield using SAR images and PS-InSAR technique. The results showed that the differential vertical deformation velocity was between -4 mm/y and 4 mm/y, while the differential horizontal deformation velocity was between -4 mm/y and 5 mm/y. Hotspots of differential vertical deformation were observed in the oilfield areas installed on piles. The study concluded that the Kashagan oilfield had not been significantly impacted by differential vertical and horizontal deformations.
Article
Environmental Studies
Sekela Twisa, Manfred E. Buchroithner
Article
Computer Science, Artificial Intelligence
Manfred F. Buchroithner
MULTIMODAL TECHNOLOGIES AND INTERACTION
(2019)
Article
Water Resources
Sekela Twisa, Shija Kazumba, Mathew Kurian, Manfred E. Buchroithner
Article
Green & Sustainable Science & Technology
Sekela Twisa, Mohamed Mwabumba, Mathew Kurian, Manfred F. Buchroithner