Article
Engineering, Marine
Zi-Lu Ouyang, Gang Chen, Zao-Jian Zou
Summary: A fast and accurate nonparametric modeling method, based on local Gaussian process regression (LGPR), is proposed for the identification modeling and prediction of ship maneuvering motion. The training dataset is automatically divided into clusters using the k-means algorithm. Local nonparametric models are identified based on the data in each cluster, reducing the computational cost compared to the classic Gaussian process regression (CGPR). Experimental data from the KVLCC2 tanker and an unmanned surface vehicle (USV) are used to identify the models, and the predictions of maneuvers not included in the training data show that LGPR has higher computational efficiency with acceptable accuracy.
Article
Construction & Building Technology
Smita Kaloni, Ghanapriya Singh, Prashant Tiwari
Summary: This paper introduces a method for nonparametric damage detection using a non-linear signal processing tool and artificial intelligence-based methodology, utilizing local mean decomposition and local gravitation clustering to extract and classify multidimensional damage features. The effectiveness of the method in damage identification is demonstrated through experiments, and its efficiency is compared with existing clustering methods.
JOURNAL OF BUILDING ENGINEERING
(2021)
Article
Computer Science, Information Systems
Zhiguo Long, Yang Gao, Hua Meng, Yuqin Yao, Tianrui Li
Summary: The article presents a clustering algorithm based on fast search and find of density peaks (DPC), which can achieve highly accurate clustering results when restricted to local neighborhoods. The proposed algorithm captures latent structures in data using family trees and incorporates both distance information and tree structure in the similarity measure. The algorithm outperforms several prominent clustering algorithms in real-world and synthetic datasets.
INFORMATION SCIENCES
(2022)
Article
Energy & Fuels
Daobo Yan, Yi Xiong, Zhihong Zhan, Xiaohong Liao, Fangchao Ke, Hailiang Lu, Yulun Ren, Shuang Liao, Lipin Sun, Qixin Wang
Summary: This paper presents a nonparametric method for selecting credit evaluation indicators under unknown index distribution, and through two rounds of screening and analysis, 18 credit characteristics suitable for power market evaluation are identified, significantly improving evaluation accuracy.
Article
Computer Science, Information Systems
Sirisup Laohakiat, Vera Sa-ing
Summary: The FIDC is an incremental density-based clustering framework that utilizes a one-pass scheme to effectively process large datasets with reduced computation time and memory usage. By employing fuzzy local clustering and a modified valley seeking algorithm, FIDC improves clustering performance and simplifies parameter selection process.
INFORMATION SCIENCES
(2021)
Article
Computer Science, Information Systems
Tao Zhang, MengChu Zhou, Xiwang Guo, Liang Qi, Abdullah Abusorrah
Summary: With the rapid development of the Internet of Things, there is a need for efficient data mining techniques to handle the vast amount of data generated. Clustering is a crucial method for discovering patterns in IoT data, and the proposed DAC algorithm tackles the challenges of finding arbitrary-shaped clusters and noise points without prior knowledge of the number of clusters. The algorithm's ability to automatically determine parameters sets it apart from other existing algorithms in the field.
IEEE INTERNET OF THINGS JOURNAL
(2022)
Article
Computer Science, Information Systems
Ren-Long Zhang, Xiao-Hong Liu
Summary: With the development of machine learning algorithm and fuzzy theory, the fuzzy clustering algorithm based on time series has gained attention. The proposed multivariate fuzzy time series clustering method based on Slacks Based Measure (MFTS-SBM) can handle fuzziness and uncertainty, and also considers the correlation of data attributes. It is important for data decision making.
Article
Computer Science, Artificial Intelligence
Wuning Tong, Yuping Wang, Delong Liu
Summary: In this study, a novel clustering algorithm LDPI based on local-density peaks is proposed to address the challenging problem of imbalanced data clustering. The algorithm has advantages such as not requiring input parameters, automatically determining cluster centers and number of clusters, and being suitable for imbalanced datasets and datasets with arbitrary shapes and distributions.
IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING
(2023)
Article
Mathematics
Mantas Lukauskas, Tomas Ruzgas
Summary: This research presents a new density clustering method based on the modified inversion formula density estimation, which divides data into different classes by discovering the internal structure of data set objects and their relationship. The method performs well, but currently can only handle low-dimensional data.
Article
Multidisciplinary Sciences
Baicheng Lyu, Wenhua Wu, Zhiqiang Hu
Summary: In this paper, a bidirectional clustering algorithm based on local density (BCALoD) is proposed, which automatically determines the number of clusters, is more sensitive to small clusters, reduces adjusted parameters, and improves clustering performance. The clustering ability of BCALoD is verified through randomly generated datasets and city light satellite images.
SCIENTIFIC REPORTS
(2021)
Article
Mathematics
Mantas Lukauskas, Tomas Ruzgas
Summary: Unsupervised learning, especially clustering methods, has numerous applications with a focus on finding hidden relationships between individual observations. This paper presents an extension to the clustering method using modified inversion formula density estimation, overcoming previous limitations and yielding improved results in higher dimensions. Comparative data analysis using over 20 data sets confirms the effectiveness of the developed method improvement. The new extended method outperforms popular data clustering methods, even approaching the accuracy of the best models, and shows positive impact on clustering results.
Article
Computer Science, Information Systems
Xiaofeng Zhang, Yujuan Sun, Hui Liu, Zhongjun Hou, Feng Zhao, Caiming Zhang
Summary: This paper proposes an improved image segmentation schema and presents two improved clustering algorithms that consider self-similarity and back projection simultaneously to enhance robustness, balancing noise restraining and detail retention in segmentation of complex images.
INFORMATION SCIENCES
(2021)
Article
Multidisciplinary Sciences
Jun-Lin Lin
Summary: This study examines different definitions for calculating the local density of data points in density-based clustering, proposing a canonical form to unify these definitions. With the canonical form, the advantages and disadvantages of existing definitions can be better explored, leading to the derivation of new definitions for local density.
Article
Automation & Control Systems
Wooseok Ha, Kimon Fountoulakis, Michael W. Mahoney
Summary: This paper analyzes the performance of the l(1)-regularized PageRank method for recovering a single target cluster and demonstrates its state-of-the-art performance on real graphs. The monotonic solution path allows for the application of the forward stagewise algorithm to approximate the entire solution path in a running time independent of the size of the whole graph. Moreover, the equivalence between l(1)-regularized PageRank and approximate personalized PageRank is established, leading to similar results for the latter method.
JOURNAL OF MACHINE LEARNING RESEARCH
(2021)
Article
Environmental Sciences
Haimiao Ge, Liguo Wang, Haizhu Pan, Yuexia Zhu, Xiaoyu Zhao, Moqi Liu
Summary: This paper proposes an improved affinity propagation algorithm based on complex wavelet structural similarity index and local outlier factor for clustering hyperspectral images. Experimental results show that the proposed method can improve the performance of traditional affinity propagation and provide competitive clustering results.