Ensemble-based clustering of large probabilistic graphs using neighborhood and distance metric learning

Danesh, Malihe; Dorrigiv, Morteza

doi:10.1007/s11227-020-03429-1

Cited by 5 publications

(3 citation statements)

References 38 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Ensemble clustering aims to produce clustering with higher quality and robustness than any of the many basic clustering algorithms by generating multiple cluster graphs of input graph and then combining their results efficiently (Ji et al, 2021). In this regard, the ENCND algorithm (Danesh et al, 2021) used the ensemble clustering method to partition the probabilistic graphs.…”

Section: Taxonomy Of Probabilistic Graph Clusteringmentioning

confidence: 99%

A survey of clustering large probabilistic graphs: Techniques, evaluations, and applications

Danesh

Dorrigiv

2023

Expert Systems

Self Cite

View full text Add to dashboard Cite

Given the growth of uncertainty in the real world, analysing probabilistic graphs is crucial. Clustering is one of the most fundamental methods of mining probabilistic graphs to discover the hidden patterns in them. This survey examines an extensive and organized analysis of the clustering techniques of large probabilistic graphs proposed in the literature. First, the definition of probabilistic graphs and modelling them are introduced. Second, the clustering of such graphs and their challenges, such as uncertainty of edges, high dimensions, and the impossibility of applying certain graph clustering techniques directly, are expressed. Then, a taxonomy of clustering approaches is discussed in two main categories: threshold-based and possible worlds-based methods. The techniques presented in each category are explained and examined. Here, these methods are evaluated on real datasets, and their performance is compared with each other. Finally, the survey is summarized by describing some of the applications of probabilistic graph clustering and future research directions.clustering, possible worlds-based methods, probabilistic graph, threshold-based methods | INTRODUCTIONGraphs are powerful tools for modelling large interconnected data networks that are highly developed in the community and the surrounding natural world. Social networks, traffic networks, genome databases, and knowledge graphs are samples of common graphs in which entities are considered nodes, and the connections between them are considered the edges. However, the uncertainty in graph data is an undeniable problem created for various reasons such as noise in measurement methods (Aggarwal & Yu, 2009), vague and inaccurate information sources (Cheng et al., 2015), lack of the required accurate information (Zhu et al., 2015), inference and prediction models (Adar & R'e, 2007), and protecting private information (Boldi et al., 2012). In these cases, the data is represented as an uncertain or probabilistic graph in which uncertainty is assigned to each graph component as a probabilistic value. It means that the probabilistic graph can be due to probabilistic nodes, edges, attributes, or a combination of them (Khan et al., 2018).However, most of the existing probabilistic graphs were formed due to probabilistic edges. Each probabilistic edge indicates the possibility of a connection between two corresponding nodes. For example, in protein interactive networks, the interaction between the two proteins is inferred experimentally and is a probabilistic value. Alternatively, in social networks, complex relationships between users, such as trust or influence, are not directly available and are indirectly inferred from user activity (Li et al., 2021). Therefore, most recent studies in probabilistic graphs have focused on these types of graphs. Accordingly, each edge is labelled with the probability of the presence of that connection.Considering the increasing number of applications with uncertain data, mining probabilistic graphs as uncertain structured data wi...

show abstract

Section: Taxonomy Of Probabilistic Graph Clusteringmentioning

confidence: 99%

A survey of clustering large probabilistic graphs: Techniques, evaluations, and applications

Danesh

Dorrigiv

2023

Expert Systems

Self Cite

View full text Add to dashboard Cite

show abstract

“…The larger the V pc , the better the clustering effect, and the smaller the V pe , the stronger the clustering effect. At the same time, the performance of the algorithm is evaluated by two clustering validity functions Davies-Bouldin index (DBI) and Dunn index (DI) [30,31]. The quantitative evaluation results are listed in Tab.…”

Section: Matching and Fusing Feature Blocksmentioning

confidence: 99%

Underwater Terrain Image Stitching Based on Spatial Gradient Feature Block

Wang¹,

Li²,

Wang³

et al. 2022

Computers, Materials &Amp; Continua

View full text Add to dashboard Cite

At present, underwater terrain images are all strip-shaped small fragment images preprocessed by the side-scan sonar imaging system. However, the processed underwater terrain images have inconspicuous and few feature points. In order to better realize the stitching of underwater terrain images and solve the problems of slow traditional image stitching speed, we proposed an improved algorithm for underwater terrain image stitching based on spatial gradient feature block. First, the spatial gradient fuzzy C-Means algorithm is used to divide the underwater terrain image into feature blocks with the fusion of spatial gradient information. The accelerated-KAZE (A-KAZE) algorithm is used to combine the feature block information to match the reference image and the target image. Then, the random sample consensus (RANSAC) is applied to optimize the matching results. Finally, image fusion is performed with the global homography and the optimal seam-line method to improve the accuracy of image overlay fusion. The experimental results show that the proposed method in this paper effectively divides images into feature blocks by combining spatial information and gradient information, which not only solves the problem of stitching failure of underwater terrain images due to unobvious features, and further reduces the sensitivity to noise, but also effectively reduces the iterative calculation in the feature point matching process of the traditional method, and improves the stitching speed. Ghosting and shape warping are significantly eliminated by re-optimizing the overlap of the image.

show abstract

“…In real-world complications diverse applications of clustering are revised e.g. in Danesh, Dorrigiv & Yaghmaee (2020). The procedure is effectively cast-off e.g.…”

Section: Introductionmentioning

confidence: 99%

A new affinity matrix weighted k-nearest neighbors graph to improve spectral clustering accuracy

Ahmed

Saeed

Paul

et al. 2021

PeerJ Computer Science

View full text Add to dashboard Cite

Researchers have thought about clustering approaches that incorporate traditional clustering methods and deep learning techniques. These approaches normally boost the performance of clustering. Getting knowledge from large data-sets is quite an interesting task. In this case, we use some dimensionality reduction and clustering techniques. Spectral clustering is gaining popularity recently because of its performance. Lately, numerous techniques have been introduced to boost spectral clustering performance. One of the most significant part of these techniques is to construct a similarity graph. We introduced weighted k-nearest neighbors technique for the construction of similarity graph. Using this new metric for the construction of affinity matrix, we achieved good results as we tested it both on real and artificial data-sets.

show abstract

Ensemble-based clustering of large probabilistic graphs using neighborhood and distance metric learning

Cited by 5 publications

References 38 publications

A survey of clustering large probabilistic graphs: Techniques, evaluations, and applications

A survey of clustering large probabilistic graphs: Techniques, evaluations, and applications

Underwater Terrain Image Stitching Based on Spatial Gradient Feature Block

A new affinity matrix weighted k-nearest neighbors graph to improve spectral clustering accuracy

Contact Info

Product

Resources

About