Article
Environmental Sciences
Haibin Yan, David Z. Zhu, Mark R. Loewen, Wenming Zhang, Shuntian Liang, Sherif Ahmed, Bert van Duin, Khizar Mahmood, Stacey Zhao
Summary: Understanding the impact of rainfall characteristics on urban stormwater quality is crucial. A data mining framework was proposed to study this relationship and a rainfall type-based calibration approach was developed to improve water quality models. The results showed that antecedent dry days, average rainfall intensity, and rainfall duration were key factors affecting stormwater quality. The K-means clustering method effectively classified rainfall events into representative types. The rainfall type-based calibration approach significantly improved model accuracy.
SCIENCE OF THE TOTAL ENVIRONMENT
(2023)
Article
Green & Sustainable Science & Technology
Sabrine Jemai, Amjad Kallel, Belgacem Agoubi, Habib Abida
Summary: This study examines the spatial and temporal rainfall variability in Gabes Catchment (southeastern Tunisia) from 1977 to 2015. Multiple statistical tools were used to characterize spatial variability of rainfall and develop a multiple regression model to predict precipitation of the different stations in the catchment area. The model shows acceptable efficiency with an absolute prediction error of approximately 87%.
ENVIRONMENT DEVELOPMENT AND SUSTAINABILITY
(2022)
Article
Engineering, Environmental
Jeremy Rohmer, Charlie Sire, Sophie Lecacheux, Deborah Idier, Rodrigo Pedreros
Summary: Metamodelling techniques can help overcome the computational burden of numerical hydrodynamic models for fast prediction of marine flooding indicators. A commonly-used approach is to reduce the dimensionality of flood maps using principal component analysis and build independent metamodels for each latent variable. However, considering the dependence structure of latent variables and clustering in their space can significantly improve the predictive performance. Using a kriging metamodel specifically designed for vector-valued variables can further enhance predictability, given a sufficient number of training samples and careful selection of clusters.
STOCHASTIC ENVIRONMENTAL RESEARCH AND RISK ASSESSMENT
(2023)
Article
Chemistry, Analytical
Carollina de Melo Molinari Ortiz Antunes, Frederico Luis Felipe Soares, Noemi Nagata
Summary: Chemical analyses based on digital images are widely studied due to their non-invasive nature and simplicity. However, controlling instrumental and structural parameters for image acquisition is crucial for analysis repeatability and reproducibility. The high cost of accessing robust instruments is also a practical limitation. To overcome these limitations, a low-cost prototype using Raspberry Pi and multivariate tools was developed.
MICROCHEMICAL JOURNAL
(2023)
Article
Mathematics, Applied
Lynne Billard, Ahlame Douzal-Chouakria, S. Yaser Samadi
Summary: Time-series data are widely studied in machine learning and data analysis for classification and clustering. However, most existing methods do not fully utilize the time-dependency information of the data. This study proposes a new method that extends principal component analysis to cross-autocorrelation matrices at different time lags to capture the main dynamic structure of multivariate time series. Experimental results on simulated data and a sign language dataset demonstrate the effectiveness and advantages of the proposed method.
Review
Environmental Sciences
Wenzhuo Wang, Lei Chen, Chen Lin, Yong Liu, Xin Dong, Junfeng Xiong, Guowangcheng Liu, Yuhan Zhang, Jiaqi Li, Zhenyao Shen
Summary: A new framework was developed in this study to undertake source apportionment at a large-scale and ungauged catchment by integrating physically-based and surrogate models. The framework was tested in the Chaohu Lake basin, China, and achieved good matching between simulated and observed data. The study identified the significant contributions of the planting industry and ungauged catchments to pollution flux, and highlighted the influence of rainfall conditions on source apportionment results.
JOURNAL OF ENVIRONMENTAL MANAGEMENT
(2023)
Article
Chemistry, Analytical
Yusheng Lang, Lilin Zhou, Yutaka Imamura
Summary: Time-of-flight secondary ion mass spectrometry (ToF-SIMS) is an important analysis technique for gathering information from surfaces. This study proposed a new approach that treats ToF-SIMS spectra as images and uses convolutional neural network (CNN) for analysis, avoiding the challenges of descriptor generation.
ANALYTICAL CHEMISTRY
(2022)
Article
Engineering, Chemical
Soleiman Hosseinpour, Alex Martynenko
Summary: The study evaluates food quality by mapping food quality attributes into multi-dimensional fuzzy sets using PCA and subtractive clustering. The methodology shows promising results in real-time quality evaluation for shrimp batch drying, with computational time below 1 second. The data-driven algorithm has unlimited potential for real-time fuzzy control and optimization.
Article
Environmental Sciences
Salvador Gil-Guirado, Alfredo Perez-Morales, David Pino, Juan Carlos Pena, Francisco Lopez Martinez
Summary: This study focused on using high-resolution flood databases and meteorological data to relate weather patterns to flood events, classifying 3608 flood cases occurring in municipalities along the Spanish Mediterranean coast between 1960 and 2015 into 12 atmospheric synoptic patterns.
SCIENCE OF THE TOTAL ENVIRONMENT
(2022)
Article
Engineering, Multidisciplinary
Zhijiang Lou, Zedong Li, Youqing Wang, Shan Lu
Summary: This paper introduces an improved neural component analysis (INCA) method, which addresses the issue of NCA's inability to handle non-Gaussian features by proposing a new cost function based on kurtosis. It also improves the extraction of key information from process data by selecting principal components (PCs) in the original data space. Experimental results show that INCA outperforms other methods in fault detection.
Article
Water Resources
Lin Zhang, Yunzhong Shen, Qiujie Chen, Fengwei Wang
Summary: By constructing the water storage deficit index (WSDI), the extreme flood event in the Pearl River Basin during the 2014-2016 super El Nin & SIM; o period was analyzed. The results showed that the flood was mainly influenced by the super El Nin & SIM; o and enhanced by the Tropical Indian Ocean sea surface temperature anomaly. The correlation analysis revealed the underlying influence mechanisms.
JOURNAL OF HYDROLOGY-REGIONAL STUDIES
(2023)
Article
Environmental Sciences
Sumita Gayen, Ismael Vallejo Villalta, Sk Mafizul Haque
Summary: This study conducted flood risk mapping in Purba Medinipur, one of the coastal districts of West Bengal, India, by considering flood frequency and vulnerability of the people. The results identified Moyna as the highest flood risk prone block and Contai-I as the least flood prone block. The findings are important for minimizing the chances of flood-related damages and establishing effective disaster management plans.
Article
Computer Science, Artificial Intelligence
Difei Cheng, Ruihang Xu, Bo Zhang, Ruinan Jin
Summary: Density-based clustering algorithms are widely used in pattern recognition and machine learning to handle non-hyperspherical clusters and outliers. However, runtime is often dominated by time-consuming neighborhood finding and density estimation processes. This paper proposes a fast range query algorithm, FPCAP, which leverages fast principal component analysis and geometric information to accelerate density-based clustering algorithms like DBSCAN and BLOCK-DBSCAN. The proposed algorithm, shown through experiments on benchmark datasets, significantly improves computational efficiency while preserving the advantages of the original algorithms.
Article
Ecology
Michael L. Collyer, Dean C. Adams
Summary: Phylogenetically aligned component analysis (PACA) is a new ordination approach that aligns phenotypic data with phylogenetic signal, allowing visualization of trends in phylogenetic signal in multivariate data spaces. By maximizing variation in directions that describe phylogenetic signal, PACA can distinguish between weak and strong phylogenetic signals, providing a more precise description of the phylogenetic signal in studies focused on phylogenetic signal. Comparing PACA and Phy-PCA results can help determine the relative importance of phylogenetic and other signals in the data.
METHODS IN ECOLOGY AND EVOLUTION
(2021)
Article
Statistics & Probability
Christian Rohrbeck, Daniel Cooley
Summary: This paper addresses the issue of generating hazard event sets of extreme river flow for northern England and southern Scotland. Through analyzing historical extreme river flow, the paper reveals interesting connections between the extremal dependence structure and the region's topography/climate. A generative framework is introduced to model the distribution of extremal principal components for generating synthetic events, which shows good agreement with observed extreme river flow dynamics.
ANNALS OF APPLIED STATISTICS
(2023)