Rare-events classification - An approach based on genetic algorithm and voronoi tessellation

Abdul Rauf Khan; Henrik Schiøler; Mohamed Zaki; Murat Kulahci

doi:10.1007/978-3-030-04503-6_26

Rare-events classification - An approach based on genetic algorithm and voronoi tessellation

Abdul Rauf Khan^*, Henrik Schiøler, Mohamed Zaki, Murat Kulahci

^*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review

Abstract

Classification is a major constituent of the data mining tool kit. Well-known methods for classification are either built on the principle of logic or on statistical reasoning. For imbalanced and noisy cases, classification may however fail to deliver on basic data mining goals, i.e., identifying statistical dependencies in data. In this article, we propose a novel strategy for data mining based on partitioning of the feature space through Voronoi tessellation and Genetic Algorithm, where the latter is applied to solve a combinatorial optimization problem. We apply the suggested methodology to a range of classification problems of varying imbalance and noise and compare the performance of the suggested method with well-known classification methods such as (SVM, KNN, and ANN). The results obtained indicate the proposed methodology to be well suited for data mining tasks in case of highly imbalanced classes and significant noise.

Original language	English
Title of host publication	Trends and Applications in Knowledge Discovery and Data Mining - PAKDD 2018 Workshops, BDASC, BDM, ML4Cyber, PAISI, DaMEMO, Revised Selected Papers
Publisher	Springer
Publication date	1 Jan 2018
Pages	256-266
ISBN (Print)	9783030045029
DOIs	https://doi.org/10.1007/978-3-030-04503-6_26
Publication status	Published - 1 Jan 2018
Event	22nd Pacific-Asia Conference on Knowledge Discovery and Data Mining - Melbourne, Australia Duration: 3 Jun 2018 → 3 Jun 2018 Conference number: 22

Conference

Conference	22nd Pacific-Asia Conference on Knowledge Discovery and Data Mining
Number	22
Country/Territory	Australia
City	Melbourne
Period	03/06/2018 → 03/06/2018

Series	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	11154
ISSN	0302-9743

Keywords

Classification
Data mining
Genetic algorithm
Imbalance
Noisy data
Voronoi tessellation

Access to Document

10.1007/978-3-030-04503-6_26

OpenUrl availability

Full text

Cite this

Khan, A. R., Schiøler, H., Zaki, M., & Kulahci, M. (2018). Rare-events classification - An approach based on genetic algorithm and voronoi tessellation. In Trends and Applications in Knowledge Discovery and Data Mining - PAKDD 2018 Workshops, BDASC, BDM, ML4Cyber, PAISI, DaMEMO, Revised Selected Papers (pp. 256-266). Springer. https://doi.org/10.1007/978-3-030-04503-6_26

Khan, Abdul Rauf ; Schiøler, Henrik ; Zaki, Mohamed et al. / Rare-events classification - An approach based on genetic algorithm and voronoi tessellation. Trends and Applications in Knowledge Discovery and Data Mining - PAKDD 2018 Workshops, BDASC, BDM, ML4Cyber, PAISI, DaMEMO, Revised Selected Papers. Springer, 2018. pp. 256-266 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), Vol. 11154 ).

@inproceedings{618a8807a0d3496498b8099880f8bce1,

title = "Rare-events classification - An approach based on genetic algorithm and voronoi tessellation",

abstract = "Classification is a major constituent of the data mining tool kit. Well-known methods for classification are either built on the principle of logic or on statistical reasoning. For imbalanced and noisy cases, classification may however fail to deliver on basic data mining goals, i.e., identifying statistical dependencies in data. In this article, we propose a novel strategy for data mining based on partitioning of the feature space through Voronoi tessellation and Genetic Algorithm, where the latter is applied to solve a combinatorial optimization problem. We apply the suggested methodology to a range of classification problems of varying imbalance and noise and compare the performance of the suggested method with well-known classification methods such as (SVM, KNN, and ANN). The results obtained indicate the proposed methodology to be well suited for data mining tasks in case of highly imbalanced classes and significant noise.",

keywords = "Classification, Data mining, Genetic algorithm, Imbalance, Noisy data, Voronoi tessellation",

author = "Khan, \{Abdul Rauf\} and Henrik Schi{\o}ler and Mohamed Zaki and Murat Kulahci",

year = "2018",

month = jan,

day = "1",

doi = "10.1007/978-3-030-04503-6\_26",

language = "English",

isbn = "9783030045029",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer",

pages = "256--266",

booktitle = "Trends and Applications in Knowledge Discovery and Data Mining - PAKDD 2018 Workshops, BDASC, BDM, ML4Cyber, PAISI, DaMEMO, Revised Selected Papers",

note = "22nd Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2018 ; Conference date: 03-06-2018 Through 03-06-2018",

}

Khan, AR, Schiøler, H, Zaki, M & Kulahci, M 2018, Rare-events classification - An approach based on genetic algorithm and voronoi tessellation. in Trends and Applications in Knowledge Discovery and Data Mining - PAKDD 2018 Workshops, BDASC, BDM, ML4Cyber, PAISI, DaMEMO, Revised Selected Papers. Springer, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 11154 , pp. 256-266, 22nd Pacific-Asia Conference on Knowledge Discovery and Data Mining, Melbourne, Victoria, Australia, 03/06/2018. https://doi.org/10.1007/978-3-030-04503-6_26

Rare-events classification - An approach based on genetic algorithm and voronoi tessellation. / Khan, Abdul Rauf; Schiøler, Henrik; Zaki, Mohamed et al.
Trends and Applications in Knowledge Discovery and Data Mining - PAKDD 2018 Workshops, BDASC, BDM, ML4Cyber, PAISI, DaMEMO, Revised Selected Papers. Springer, 2018. p. 256-266 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), Vol. 11154 ).

Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review

TY - GEN

T1 - Rare-events classification - An approach based on genetic algorithm and voronoi tessellation

AU - Khan, Abdul Rauf

AU - Schiøler, Henrik

AU - Zaki, Mohamed

AU - Kulahci, Murat

N1 - Conference code: 22

PY - 2018/1/1

Y1 - 2018/1/1

N2 - Classification is a major constituent of the data mining tool kit. Well-known methods for classification are either built on the principle of logic or on statistical reasoning. For imbalanced and noisy cases, classification may however fail to deliver on basic data mining goals, i.e., identifying statistical dependencies in data. In this article, we propose a novel strategy for data mining based on partitioning of the feature space through Voronoi tessellation and Genetic Algorithm, where the latter is applied to solve a combinatorial optimization problem. We apply the suggested methodology to a range of classification problems of varying imbalance and noise and compare the performance of the suggested method with well-known classification methods such as (SVM, KNN, and ANN). The results obtained indicate the proposed methodology to be well suited for data mining tasks in case of highly imbalanced classes and significant noise.

AB - Classification is a major constituent of the data mining tool kit. Well-known methods for classification are either built on the principle of logic or on statistical reasoning. For imbalanced and noisy cases, classification may however fail to deliver on basic data mining goals, i.e., identifying statistical dependencies in data. In this article, we propose a novel strategy for data mining based on partitioning of the feature space through Voronoi tessellation and Genetic Algorithm, where the latter is applied to solve a combinatorial optimization problem. We apply the suggested methodology to a range of classification problems of varying imbalance and noise and compare the performance of the suggested method with well-known classification methods such as (SVM, KNN, and ANN). The results obtained indicate the proposed methodology to be well suited for data mining tasks in case of highly imbalanced classes and significant noise.

KW - Classification

KW - Data mining

KW - Genetic algorithm

KW - Imbalance

KW - Noisy data

KW - Voronoi tessellation

U2 - 10.1007/978-3-030-04503-6_26

DO - 10.1007/978-3-030-04503-6_26

M3 - Article in proceedings

AN - SCOPUS:85059055751

SN - 9783030045029

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 256

EP - 266

BT - Trends and Applications in Knowledge Discovery and Data Mining - PAKDD 2018 Workshops, BDASC, BDM, ML4Cyber, PAISI, DaMEMO, Revised Selected Papers

PB - Springer

T2 - 22nd Pacific-Asia Conference on Knowledge Discovery and Data Mining

Y2 - 3 June 2018 through 3 June 2018

ER -

Khan AR, Schiøler H, Zaki M, Kulahci M. Rare-events classification - An approach based on genetic algorithm and voronoi tessellation. In Trends and Applications in Knowledge Discovery and Data Mining - PAKDD 2018 Workshops, BDASC, BDM, ML4Cyber, PAISI, DaMEMO, Revised Selected Papers. Springer. 2018. p. 256-266. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), Vol. 11154 ). doi: 10.1007/978-3-030-04503-6_26

Rare-events classification - An approach based on genetic algorithm and voronoi tessellation

Abstract

Conference

Keywords

Access to Document

OpenUrl availability

Fingerprint

Cite this