A novel scalable multi-class ROC for effective visualization and computation

Md Rafiul Hassan; Kotagiri Ramamohanarao; Chandan Karmakar; M. Maruf Hossain; James Bailey

doi:10.1007/978-3-642-13657-3_14

A novel scalable multi-class ROC for effective visualization and computation

Md Rafiul Hassan, Kotagiri Ramamohanarao, Chandan Karmakar, M. Maruf Hossain, James Bailey

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

22 Scopus citations

Abstract

This paper introduces a new cost function for evaluating the multi-class classifier. The new cost function facilitates both a way to visualize the performance (expected cost) of the multi-class classifier and a summary of the misclassification costs. This function overcomes the limitations of ROC in not being able to represent the classifier performance graphically when there are more than two classes. Here we present a new scalable method for producing a scalar measurement that is used to compare the performance of the multi-class classifier. We mathematically demonstrate that our technique can capture small variations in classifier performance.

Original language	English
Title of host publication	Advances in Knowledge Discovery and Data Mining - 14th Pacific-Asia Conference, PAKDD 2010, Proceedings
Pages	107-120
Number of pages	14
Edition	PART 1
DOIs	https://doi.org/10.1007/978-3-642-13657-3_14
State	Published - 2010
Externally published	Yes
Event	14th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2010 - Hyderabad, India Duration: 21 Jun 2010 → 24 Jun 2010

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Number	PART 1
Volume	6118 LNAI
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	14th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2010
Country/Territory	India
City	Hyderabad
Period	21/06/10 → 24/06/10

Keywords

Classifier evaluation
Cost-function
Multi-class
Receiver operating characteristics

ASJC Scopus subject areas

Theoretical Computer Science
General Computer Science

Access to Document

10.1007/978-3-642-13657-3_14

Cite this

Hassan, M. R., Ramamohanarao, K., Karmakar, C., Hossain, M. M., & Bailey, J. (2010). A novel scalable multi-class ROC for effective visualization and computation. In Advances in Knowledge Discovery and Data Mining - 14th Pacific-Asia Conference, PAKDD 2010, Proceedings (PART 1 ed., pp. 107-120). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 6118 LNAI, No. PART 1). https://doi.org/10.1007/978-3-642-13657-3_14

Hassan, Md Rafiul ; Ramamohanarao, Kotagiri ; Karmakar, Chandan et al. / A novel scalable multi-class ROC for effective visualization and computation. Advances in Knowledge Discovery and Data Mining - 14th Pacific-Asia Conference, PAKDD 2010, Proceedings. PART 1. ed. 2010. pp. 107-120 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 1).

@inproceedings{c7c6ad0d44d449e98ca5a2ed0f215ab2,

title = "A novel scalable multi-class ROC for effective visualization and computation",

abstract = "This paper introduces a new cost function for evaluating the multi-class classifier. The new cost function facilitates both a way to visualize the performance (expected cost) of the multi-class classifier and a summary of the misclassification costs. This function overcomes the limitations of ROC in not being able to represent the classifier performance graphically when there are more than two classes. Here we present a new scalable method for producing a scalar measurement that is used to compare the performance of the multi-class classifier. We mathematically demonstrate that our technique can capture small variations in classifier performance.",

keywords = "Classifier evaluation, Cost-function, Multi-class, Receiver operating characteristics",

author = "Hassan, \{Md Rafiul\} and Kotagiri Ramamohanarao and Chandan Karmakar and Hossain, \{M. Maruf\} and James Bailey",

year = "2010",

doi = "10.1007/978-3-642-13657-3\_14",

language = "English",

isbn = "3642136567",

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

number = "PART 1",

pages = "107--120",

booktitle = "Advances in Knowledge Discovery and Data Mining - 14th Pacific-Asia Conference, PAKDD 2010, Proceedings",

edition = "PART 1",

note = "14th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2010 ; Conference date: 21-06-2010 Through 24-06-2010",

}

Hassan, MR, Ramamohanarao, K, Karmakar, C, Hossain, MM & Bailey, J 2010, A novel scalable multi-class ROC for effective visualization and computation. in Advances in Knowledge Discovery and Data Mining - 14th Pacific-Asia Conference, PAKDD 2010, Proceedings. PART 1 edn, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), no. PART 1, vol. 6118 LNAI, pp. 107-120, 14th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2010, Hyderabad, India, 21/06/10. https://doi.org/10.1007/978-3-642-13657-3_14

A novel scalable multi-class ROC for effective visualization and computation. / Hassan, Md Rafiul; Ramamohanarao, Kotagiri; Karmakar, Chandan et al.
Advances in Knowledge Discovery and Data Mining - 14th Pacific-Asia Conference, PAKDD 2010, Proceedings. PART 1. ed. 2010. p. 107-120 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 6118 LNAI, No. PART 1).

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

TY - GEN

T1 - A novel scalable multi-class ROC for effective visualization and computation

AU - Hassan, Md Rafiul

AU - Ramamohanarao, Kotagiri

AU - Karmakar, Chandan

AU - Hossain, M. Maruf

AU - Bailey, James

PY - 2010

Y1 - 2010

N2 - This paper introduces a new cost function for evaluating the multi-class classifier. The new cost function facilitates both a way to visualize the performance (expected cost) of the multi-class classifier and a summary of the misclassification costs. This function overcomes the limitations of ROC in not being able to represent the classifier performance graphically when there are more than two classes. Here we present a new scalable method for producing a scalar measurement that is used to compare the performance of the multi-class classifier. We mathematically demonstrate that our technique can capture small variations in classifier performance.

AB - This paper introduces a new cost function for evaluating the multi-class classifier. The new cost function facilitates both a way to visualize the performance (expected cost) of the multi-class classifier and a summary of the misclassification costs. This function overcomes the limitations of ROC in not being able to represent the classifier performance graphically when there are more than two classes. Here we present a new scalable method for producing a scalar measurement that is used to compare the performance of the multi-class classifier. We mathematically demonstrate that our technique can capture small variations in classifier performance.

KW - Classifier evaluation

KW - Cost-function

KW - Multi-class

KW - Receiver operating characteristics

UR - https://www.scopus.com/pages/publications/79956301925

U2 - 10.1007/978-3-642-13657-3_14

DO - 10.1007/978-3-642-13657-3_14

M3 - Conference contribution

AN - SCOPUS:79956301925

SN - 3642136567

SN - 9783642136566

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

SP - 107

EP - 120

BT - Advances in Knowledge Discovery and Data Mining - 14th Pacific-Asia Conference, PAKDD 2010, Proceedings

T2 - 14th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2010

Y2 - 21 June 2010 through 24 June 2010

ER -

Hassan MR, Ramamohanarao K, Karmakar C, Hossain MM, Bailey J. A novel scalable multi-class ROC for effective visualization and computation. In Advances in Knowledge Discovery and Data Mining - 14th Pacific-Asia Conference, PAKDD 2010, Proceedings. PART 1 ed. 2010. p. 107-120. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 1). doi: 10.1007/978-3-642-13657-3_14

A novel scalable multi-class ROC for effective visualization and computation

Abstract

Publication series

Conference

Keywords

ASJC Scopus subject areas

Access to Document

Fingerprint

Cite this