Abstract

Support vector machines (SVMs) are supervised learning models traditionally employed for classification and regression analysis. In classification analysis, a set of training data is chosen, and each instance in the training data is assigned a categorical class. An SVM then constructs a model based on a separating plane that maximizes the margin between different classes. Despite being one of the most popular classification models because of its strong performance empirically, understanding the knowledge captured in an SVM remains difficult. SVMs are typically applied in a black-box manner where the details of parameter tuning, training, and even the final constructed model are hidden from the users. This is natural since these details are often complex and difficult to understand without proper visualization tools. However, such an approach often brings about various problems including trial-and-error tuning and suspicious users who are forced to trust these models blindly. The contribution of this paper is a visual analysis approach for building SVMs in an open-box manner. Our goal is to improve an analyst’s understanding of the SVM modeling process through a suite of visualization techniques that allow users to have full interactive visual control over the entire SVM training process. Our visual exploration tools have been developed to enable intuitive parameter tuning, training data manipulation, and rule extraction as part of the SVM training process. To demonstrate the efficacy of our approach, we conduct a case study using a real-world robot control dataset.

Original languageEnglish (US)
Pages (from-to)161-175
Number of pages15
JournalComputational Visual Media
Volume3
Issue number2
DOIs
StatePublished - Jun 1 2017

Fingerprint

Support vector machines
Tuning
Visualization
Supervised learning
Regression analysis
Robots

Keywords

  • high-dimensional visualization
  • rule extraction
  • support vector machines (SVMs)
  • visual analysis
  • visual classification

ASJC Scopus subject areas

  • Computer Graphics and Computer-Aided Design
  • Computer Vision and Pattern Recognition
  • Artificial Intelligence

Cite this

EasySVM : A visual analysis approach for open-box support vector machines. / Ma, Yuxin; Chen, Wei; Ma, Xiaohong; Xu, Jiayi; Huang, Xinxin; Maciejewski, Ross; Tung, Anthony K.H.

In: Computational Visual Media, Vol. 3, No. 2, 01.06.2017, p. 161-175.

Research output: Contribution to journalArticle

Ma, Yuxin ; Chen, Wei ; Ma, Xiaohong ; Xu, Jiayi ; Huang, Xinxin ; Maciejewski, Ross ; Tung, Anthony K.H. / EasySVM : A visual analysis approach for open-box support vector machines. In: Computational Visual Media. 2017 ; Vol. 3, No. 2. pp. 161-175.
@article{0b196d1bb30e4f4ba8d5b59c9c18de1c,
title = "EasySVM: A visual analysis approach for open-box support vector machines",
abstract = "Support vector machines (SVMs) are supervised learning models traditionally employed for classification and regression analysis. In classification analysis, a set of training data is chosen, and each instance in the training data is assigned a categorical class. An SVM then constructs a model based on a separating plane that maximizes the margin between different classes. Despite being one of the most popular classification models because of its strong performance empirically, understanding the knowledge captured in an SVM remains difficult. SVMs are typically applied in a black-box manner where the details of parameter tuning, training, and even the final constructed model are hidden from the users. This is natural since these details are often complex and difficult to understand without proper visualization tools. However, such an approach often brings about various problems including trial-and-error tuning and suspicious users who are forced to trust these models blindly. The contribution of this paper is a visual analysis approach for building SVMs in an open-box manner. Our goal is to improve an analyst’s understanding of the SVM modeling process through a suite of visualization techniques that allow users to have full interactive visual control over the entire SVM training process. Our visual exploration tools have been developed to enable intuitive parameter tuning, training data manipulation, and rule extraction as part of the SVM training process. To demonstrate the efficacy of our approach, we conduct a case study using a real-world robot control dataset.",
keywords = "high-dimensional visualization, rule extraction, support vector machines (SVMs), visual analysis, visual classification",
author = "Yuxin Ma and Wei Chen and Xiaohong Ma and Jiayi Xu and Xinxin Huang and Ross Maciejewski and Tung, {Anthony K.H.}",
year = "2017",
month = "6",
day = "1",
doi = "10.1007/s41095-017-0077-5",
language = "English (US)",
volume = "3",
pages = "161--175",
journal = "Computational Visual Media",
issn = "2096-0433",
publisher = "Tsinghua University Press",
number = "2",

}

TY - JOUR

T1 - EasySVM

T2 - A visual analysis approach for open-box support vector machines

AU - Ma, Yuxin

AU - Chen, Wei

AU - Ma, Xiaohong

AU - Xu, Jiayi

AU - Huang, Xinxin

AU - Maciejewski, Ross

AU - Tung, Anthony K.H.

PY - 2017/6/1

Y1 - 2017/6/1

N2 - Support vector machines (SVMs) are supervised learning models traditionally employed for classification and regression analysis. In classification analysis, a set of training data is chosen, and each instance in the training data is assigned a categorical class. An SVM then constructs a model based on a separating plane that maximizes the margin between different classes. Despite being one of the most popular classification models because of its strong performance empirically, understanding the knowledge captured in an SVM remains difficult. SVMs are typically applied in a black-box manner where the details of parameter tuning, training, and even the final constructed model are hidden from the users. This is natural since these details are often complex and difficult to understand without proper visualization tools. However, such an approach often brings about various problems including trial-and-error tuning and suspicious users who are forced to trust these models blindly. The contribution of this paper is a visual analysis approach for building SVMs in an open-box manner. Our goal is to improve an analyst’s understanding of the SVM modeling process through a suite of visualization techniques that allow users to have full interactive visual control over the entire SVM training process. Our visual exploration tools have been developed to enable intuitive parameter tuning, training data manipulation, and rule extraction as part of the SVM training process. To demonstrate the efficacy of our approach, we conduct a case study using a real-world robot control dataset.

AB - Support vector machines (SVMs) are supervised learning models traditionally employed for classification and regression analysis. In classification analysis, a set of training data is chosen, and each instance in the training data is assigned a categorical class. An SVM then constructs a model based on a separating plane that maximizes the margin between different classes. Despite being one of the most popular classification models because of its strong performance empirically, understanding the knowledge captured in an SVM remains difficult. SVMs are typically applied in a black-box manner where the details of parameter tuning, training, and even the final constructed model are hidden from the users. This is natural since these details are often complex and difficult to understand without proper visualization tools. However, such an approach often brings about various problems including trial-and-error tuning and suspicious users who are forced to trust these models blindly. The contribution of this paper is a visual analysis approach for building SVMs in an open-box manner. Our goal is to improve an analyst’s understanding of the SVM modeling process through a suite of visualization techniques that allow users to have full interactive visual control over the entire SVM training process. Our visual exploration tools have been developed to enable intuitive parameter tuning, training data manipulation, and rule extraction as part of the SVM training process. To demonstrate the efficacy of our approach, we conduct a case study using a real-world robot control dataset.

KW - high-dimensional visualization

KW - rule extraction

KW - support vector machines (SVMs)

KW - visual analysis

KW - visual classification

UR - http://www.scopus.com/inward/record.url?scp=85046766170&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85046766170&partnerID=8YFLogxK

U2 - 10.1007/s41095-017-0077-5

DO - 10.1007/s41095-017-0077-5

M3 - Article

VL - 3

SP - 161

EP - 175

JO - Computational Visual Media

JF - Computational Visual Media

SN - 2096-0433

IS - 2

ER -