Hyper-ellipsoidal clustering technique for evolving data stream

Rehman, Muhammad Zia-ur; Li, Tianrui; Yang, Yan; Wang, Hongjun

doi:10.1016/j.knosys.2013.11.022

Cited by 26 publications

(13 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For datasets with more than 500 fields, we use SOFM, which can be viewed as an independent parallel process of AdaHS, to obtain 500 cluster centers and use the centers as the representative points in (6). Previous research showed that sampling with clustering can enable Nyström method to have a much better approximation than uniform sampling [36].…”

Section: B Kernel Approximation With Nyström Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Classifying With Adaptive Hyper-Spheres: An Incremental Classifier Based on Competitive Learning

Kou

Peng

et al. 2020

IEEE Trans. Syst. Man Cybern, Syst.

View full text Add to dashboard Cite

Nowadays, datasets are always dynamic and patterns in them are changing. Instances with different labels are intertwined and often linearly inseparable, which bring new challenges to traditional learning algorithms. This paper proposes adaptive hyper-sphere (AdaHS), an adaptive incremental classifier, and its kernelized version: Nys-AdaHS. The classifier incorporates competitive training with a border zone. With adaptive hidden layer and tunable radii of hyper-spheres, AdaHS has strong capability of local learning like instance-based algorithms, but free from slow searching speed and excessive memory consumption. The experiments showed that AdaHS is robust, adaptive, and highly accurate. It is especially suitable for dynamic data in which patterns are changing, decision borders are complicated, and instances with the same label can be spherically clustered.

show abstract

Section: B Kernel Approximation With Nyström Methodsmentioning

confidence: 99%

“…One single hyper-sphere may not enclose an area whose shape is not hyper-spherical [36]. However, any shape could be enclosed as long as the number of the formed hyper-spheres…”

Section: Building Of the Dmzmentioning

confidence: 99%

Classifying With Adaptive Hyper-Spheres: An Incremental Classifier Based on Competitive Learning

Kou

Peng

et al. 2020

IEEE Trans. Syst. Man Cybern, Syst.

View full text Add to dashboard Cite

show abstract

“…The main sampling methods are based upon the distance and density concepts. A survey shows that the corresponding methods have reached a high level of maturity [39,16,50,58,20].…”

Section: Introductionmentioning

confidence: 99%

DIDES: a fast and effective sampling for clustering algorithm

Ros

Guillaume²

2016

Knowl Inf Syst

View full text Add to dashboard Cite

As clustering algorithms become more and more sophisticated to cope with current needs, large data sets of increasing complexity, sampling is likely to provide an interesting alternative. The proposal is a distancebased algorithm: the idea is to iteratively include in the sample the furthest item from all the already selected ones. Density is managed within a post-processing step, either low or high density areas are considered. The algorithm has some nice properties: insensitive to initialization, data size and noise, it is accurate according to the Rand index and avoids many distance calculations thanks to internal optimization. Moreover it is driven by only one, meaningful, parameter, called granularity, which impacts the sample size. Compared with concurrent approaches, it proved to be as powerful as the best known methods, with the lowest CPU cost.

show abstract

“…Two recently published papers proposed a more flexible assumption regarding the shape of clusters by using ellipsoidal shape instead of spherical shape as in Clustream. HECES [7] uses grid-cells to calculate the statistical summary of streaming data. The grid-cells are replaced by a hyper-ellipsoidal shape from the covariance of grid-cells.…”

Section: Literature Reviewsmentioning

confidence: 99%

“…In this set of experiments, to compare the clustering performance of the algorithms without the effect of fading data, the fading function of the micro-clusters was excluded from the algorithms. To determine the dynamic performance in clustering evolving stream, the sliding window model was adopted as in [7]. In this set of experiments, the parameters of LLDstream were set at r=0.1, np=15 for all data sets.…”

Section: Comparison With Algorithms For Streaming Datamentioning

confidence: 99%

Development of Density Based Clustering Algorithms for Streaming Data

Laohakiat

View full text Add to dashboard Cite

Streaming data has played important role in many data mining applications. Different from traditional data sets which the whole data records are available at the beginning, streaming data accumulate over time, and usually due to its continuous flow of data records, the volume of this kind of data set is usually large. Several algorithms for clustering streaming data have been designed in accordance with these restrictions. In this study, some further constraints on the characteristics of the data set are considered; in order to design density based clustering algorithms which can cluster these data sets efficiently. The designed algorithms have been tested against some state-of-the-art algorithms to determine the effectiveness of the proposed algorithms.

show abstract

Hyper-ellipsoidal clustering technique for evolving data stream

Cited by 26 publications

References 18 publications

Classifying With Adaptive Hyper-Spheres: An Incremental Classifier Based on Competitive Learning

Classifying With Adaptive Hyper-Spheres: An Incremental Classifier Based on Competitive Learning

DIDES: a fast and effective sampling for clustering algorithm

Development of Density Based Clustering Algorithms for Streaming Data

Contact Info

Product

Resources

About