An interactive machine-learning method to obtain safety information from free text

Peter Hughes; Coen Van Gulijk; Rawia El Rashidy

An interactive machine-learning method to obtain safety information from free text

Peter Hughes, Coen Van Gulijk, Rawia El Rashidy

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

Abstract

This paper describes the continued development of natural language processing (NLP) techniques to support safety management on the GB railways. The work considers machine reading techniques to obtain information from more than 800,000 free text hazard records in the Close Call System (CCS) Our research has found that the non-standard nature of the text means that standard NLP techniques yield accuracies between 0% and 60% when categorising hazard reports. To improve accuracy, a workflow was developed that uses the results from individual techniques in an iterative cycle with a human analyst. The technique, dubbed interactive learning, has achieved substantially improved accuracy of up to 98% and is currently being integrated by the GB railway industry as part of its SMS.

Original language	English
Title of host publication	Proceedings of the 29th European Safety and Reliability Conference (ESREL 2019)
Editors	Michael Beer, Enrico Zio
Pages	46-54
Number of pages	9
ISBN (Electronic)	9789811127243
Publication status	Published - Sep 2019
Event	29th European Safety and Reliability Conference - Leibniz Universität, Hannover, Germany Duration: 22 Sep 2019 → 26 Sep 2019 Conference number: 29 https://esrel2019.org/#/

Conference

Conference	29th European Safety and Reliability Conference
Abbreviated title	ESREL 2019
Country/Territory	Germany
City	Hannover
Period	22/09/19 → 26/09/19
Internet address	https://esrel2019.org/#/

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

Cite this

@inproceedings{714db834ba7843c2859f0ce2dd29f4c5,

title = "An interactive machine-learning method to obtain safety information from free text",

abstract = "This paper describes the continued development of natural language processing (NLP) techniques to support safety management on the GB railways. The work considers machine reading techniques to obtain information from more than 800,000 free text hazard records in the Close Call System (CCS) Our research has found that the non-standard nature of the text means that standard NLP techniques yield accuracies between 0% and 60% when categorising hazard reports. To improve accuracy, a workflow was developed that uses the results from individual techniques in an iterative cycle with a human analyst. The technique, dubbed interactive learning, has achieved substantially improved accuracy of up to 98% and is currently being integrated by the GB railway industry as part of its SMS.",

keywords = "close calls, natural language processing, interactive learning, text analysis",

author = "Peter Hughes and {Van Gulijk}, Coen and {El Rashidy}, Rawia",

year = "2019",

month = sep,

language = "English",

pages = "46--54",

editor = "Michael Beer and Enrico Zio",

booktitle = "Proceedings of the 29th European Safety and Reliability Conference (ESREL 2019)",

note = "29th European Safety and Reliability Conference, ESREL 2019 ; Conference date: 22-09-2019 Through 26-09-2019",

url = "https://esrel2019.org/#/",

}

An interactive machine-learning method to obtain safety information from free text. / Hughes, Peter; Van Gulijk, Coen; El Rashidy, Rawia.

Proceedings of the 29th European Safety and Reliability Conference (ESREL 2019). ed. / Michael Beer; Enrico Zio. 2019. p. 46-54.

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

TY - GEN

T1 - An interactive machine-learning method to obtain safety information from free text

AU - Hughes, Peter

AU - Van Gulijk, Coen

AU - El Rashidy, Rawia

N1 - Conference code: 29

PY - 2019/9

Y1 - 2019/9

N2 - This paper describes the continued development of natural language processing (NLP) techniques to support safety management on the GB railways. The work considers machine reading techniques to obtain information from more than 800,000 free text hazard records in the Close Call System (CCS) Our research has found that the non-standard nature of the text means that standard NLP techniques yield accuracies between 0% and 60% when categorising hazard reports. To improve accuracy, a workflow was developed that uses the results from individual techniques in an iterative cycle with a human analyst. The technique, dubbed interactive learning, has achieved substantially improved accuracy of up to 98% and is currently being integrated by the GB railway industry as part of its SMS.

AB - This paper describes the continued development of natural language processing (NLP) techniques to support safety management on the GB railways. The work considers machine reading techniques to obtain information from more than 800,000 free text hazard records in the Close Call System (CCS) Our research has found that the non-standard nature of the text means that standard NLP techniques yield accuracies between 0% and 60% when categorising hazard reports. To improve accuracy, a workflow was developed that uses the results from individual techniques in an iterative cycle with a human analyst. The technique, dubbed interactive learning, has achieved substantially improved accuracy of up to 98% and is currently being integrated by the GB railway industry as part of its SMS.

KW - close calls

KW - natural language processing

KW - interactive learning

KW - text analysis

UR - https://esrel2019.org/#/

UR - http://itekcmsonline.com/rps2prod/esrel2019/e-proceedings/index.html

M3 - Conference contribution

SP - 46

EP - 54

BT - Proceedings of the 29th European Safety and Reliability Conference (ESREL 2019)

A2 - Beer, Michael

A2 - Zio, Enrico

T2 - 29th European Safety and Reliability Conference

Y2 - 22 September 2019 through 26 September 2019

ER -

An interactive machine-learning method to obtain safety information from free text

Abstract

Conference

UN SDGs

Other files and links

Fingerprint

Cite this