Diagnostic Prediction Using Discomfort Drawings with IBTM

Cheng Zhang; Hedvig Kjellström; Carl Henrik Ek; Bo C. Bertilson

Diagnostic Prediction Using Discomfort Drawings with IBTM

Cheng Zhang
, Hedvig Kjellström
, Carl Henrik Ek
, Bo C. Bertilson

Research output: Chapter in Book/Report/Conference proceeding › Conference Contribution (Conference Proceeding)

88 Downloads (Pure)

Abstract

In this paper, we explore the possibility to apply machine learning to make diagnostic predictions using discomfort drawings. A discomfort drawing is an intuitive way for patients to express discomfort and pain related symptoms. These drawings have proven to be an effective method to collect patient data and make diagnostic decisions in real-life practice. A dataset from real-world patient cases is collected for which medical experts provide diagnostic labels. Next, we use a factorized multimodal topic model, Inter-Battery Topic Model (IBTM), to train a system that can make diagnostic predictions given an unseen discomfort drawing. The number of output diagnostic labels is determined by using mean-shift clustering on the discomfort drawing. Experimental results show reasonable predictions of diagnostic labels given an unseen discomfort drawing. Additionally, we generate synthetic discomfort drawings with IBTM given a diagnostic label, which results in typical cases of symptoms. The positive result indicates a significant potential of machine learning to be used for parts of the pain diagnostic process and to be a decision support system for physicians and other health care personnel.

Original language	English
Title of host publication	Proceedings of the 1st Machine Learning in Health Care, MLHC 2016, Los Angeles, CA, USA, August 19-20, 2016
Editors	Finale Doshi-Velez, Jim Fackler, David C. Kale, Byron C. Wallace, Jenna Wiens
Publisher	JMLR.org
Pages	226-238
Number of pages	13
Volume	56
Publication status	Published - 27 Jul 2016

Publication series

Name	JMLR Workshop and Conference Proceedings
Publisher	JMLR.org

Access to Document

Full-text PDF (accepted author manuscript)
This is the submitted manuscript. It is also available online via arXiv at https://arxiv.org/abs/1607.08206. Please refer to any applicable terms of use of the publisher.
Accepted author manuscript, 2.53 MB

http://proceedings.mlr.press/v56/Zhang16.html

Handle.net

Persistent link

Cite this

Zhang, C., Kjellström, H., Ek, C. H., & Bertilson, B. C. (2016). Diagnostic Prediction Using Discomfort Drawings with IBTM. In F. Doshi-Velez, J. Fackler, D. C. Kale, B. C. Wallace, & J. Wiens (Eds.), Proceedings of the 1st Machine Learning in Health Care, MLHC 2016, Los Angeles, CA, USA, August 19-20, 2016 (Vol. 56, pp. 226-238). (JMLR Workshop and Conference Proceedings). JMLR.org. http://proceedings.mlr.press/v56/Zhang16.html

Zhang, Cheng ; Kjellström, Hedvig ; Ek, Carl Henrik et al. / Diagnostic Prediction Using Discomfort Drawings with IBTM. Proceedings of the 1st Machine Learning in Health Care, MLHC 2016, Los Angeles, CA, USA, August 19-20, 2016. editor / Finale Doshi-Velez ; Jim Fackler ; David C. Kale ; Byron C. Wallace ; Jenna Wiens. Vol. 56 JMLR.org, 2016. pp. 226-238 (JMLR Workshop and Conference Proceedings).

@inproceedings{df044648e7d842738878da20378f52e6,

title = "Diagnostic Prediction Using Discomfort Drawings with IBTM",

abstract = "In this paper, we explore the possibility to apply machine learning to make diagnostic predictions using discomfort drawings. A discomfort drawing is an intuitive way for patients to express discomfort and pain related symptoms. These drawings have proven to be an effective method to collect patient data and make diagnostic decisions in real-life practice. A dataset from real-world patient cases is collected for which medical experts provide diagnostic labels. Next, we use a factorized multimodal topic model, Inter-Battery Topic Model (IBTM), to train a system that can make diagnostic predictions given an unseen discomfort drawing. The number of output diagnostic labels is determined by using mean-shift clustering on the discomfort drawing. Experimental results show reasonable predictions of diagnostic labels given an unseen discomfort drawing. Additionally, we generate synthetic discomfort drawings with IBTM given a diagnostic label, which results in typical cases of symptoms. The positive result indicates a significant potential of machine learning to be used for parts of the pain diagnostic process and to be a decision support system for physicians and other health care personnel. ",

author = "Cheng Zhang and Hedvig Kjellstr{\"o}m and Ek, \{Carl Henrik\} and Bertilson, \{Bo C.\}",

year = "2016",

month = jul,

day = "27",

language = "English",

volume = "56",

series = "JMLR Workshop and Conference Proceedings",

publisher = "JMLR.org",

pages = "226--238",

editor = "Finale Doshi-Velez and Jim Fackler and Kale, \{David C.\} and Wallace, \{Byron C.\} and Jenna Wiens",

booktitle = "Proceedings of the 1st Machine Learning in Health Care, MLHC 2016, Los Angeles, CA, USA, August 19-20, 2016",

}

Zhang, C, Kjellström, H, Ek, CH & Bertilson, BC 2016, Diagnostic Prediction Using Discomfort Drawings with IBTM. in F Doshi-Velez, J Fackler, DC Kale, BC Wallace & J Wiens (eds), Proceedings of the 1st Machine Learning in Health Care, MLHC 2016, Los Angeles, CA, USA, August 19-20, 2016. vol. 56, JMLR Workshop and Conference Proceedings, JMLR.org, pp. 226-238. <http://proceedings.mlr.press/v56/Zhang16.html>

Diagnostic Prediction Using Discomfort Drawings with IBTM. / Zhang, Cheng; Kjellström, Hedvig; Ek, Carl Henrik et al.
Proceedings of the 1st Machine Learning in Health Care, MLHC 2016, Los Angeles, CA, USA, August 19-20, 2016. ed. / Finale Doshi-Velez; Jim Fackler; David C. Kale; Byron C. Wallace; Jenna Wiens. Vol. 56 JMLR.org, 2016. p. 226-238 (JMLR Workshop and Conference Proceedings).

Research output: Chapter in Book/Report/Conference proceeding › Conference Contribution (Conference Proceeding)

TY - GEN

T1 - Diagnostic Prediction Using Discomfort Drawings with IBTM

AU - Zhang, Cheng

AU - Kjellström, Hedvig

AU - Ek, Carl Henrik

AU - Bertilson, Bo C.

PY - 2016/7/27

Y1 - 2016/7/27

N2 - In this paper, we explore the possibility to apply machine learning to make diagnostic predictions using discomfort drawings. A discomfort drawing is an intuitive way for patients to express discomfort and pain related symptoms. These drawings have proven to be an effective method to collect patient data and make diagnostic decisions in real-life practice. A dataset from real-world patient cases is collected for which medical experts provide diagnostic labels. Next, we use a factorized multimodal topic model, Inter-Battery Topic Model (IBTM), to train a system that can make diagnostic predictions given an unseen discomfort drawing. The number of output diagnostic labels is determined by using mean-shift clustering on the discomfort drawing. Experimental results show reasonable predictions of diagnostic labels given an unseen discomfort drawing. Additionally, we generate synthetic discomfort drawings with IBTM given a diagnostic label, which results in typical cases of symptoms. The positive result indicates a significant potential of machine learning to be used for parts of the pain diagnostic process and to be a decision support system for physicians and other health care personnel.

AB - In this paper, we explore the possibility to apply machine learning to make diagnostic predictions using discomfort drawings. A discomfort drawing is an intuitive way for patients to express discomfort and pain related symptoms. These drawings have proven to be an effective method to collect patient data and make diagnostic decisions in real-life practice. A dataset from real-world patient cases is collected for which medical experts provide diagnostic labels. Next, we use a factorized multimodal topic model, Inter-Battery Topic Model (IBTM), to train a system that can make diagnostic predictions given an unseen discomfort drawing. The number of output diagnostic labels is determined by using mean-shift clustering on the discomfort drawing. Experimental results show reasonable predictions of diagnostic labels given an unseen discomfort drawing. Additionally, we generate synthetic discomfort drawings with IBTM given a diagnostic label, which results in typical cases of symptoms. The positive result indicates a significant potential of machine learning to be used for parts of the pain diagnostic process and to be a decision support system for physicians and other health care personnel.

UR - https://arxiv.org/abs/1607.08206

M3 - Conference Contribution (Conference Proceeding)

VL - 56

T3 - JMLR Workshop and Conference Proceedings

SP - 226

EP - 238

BT - Proceedings of the 1st Machine Learning in Health Care, MLHC 2016, Los Angeles, CA, USA, August 19-20, 2016

A2 - Doshi-Velez, Finale

A2 - Fackler, Jim

A2 - Kale, David C.

A2 - Wallace, Byron C.

A2 - Wiens, Jenna

PB - JMLR.org

ER -

Diagnostic Prediction Using Discomfort Drawings with IBTM

Abstract

Publication series

Access to Document

Handle.net

Other files and links

Fingerprint

Cite this