Learning I/O Variables from Scientific Software’s User Manuals

Zedong Peng; Xuanyi Lin; Sreelekhaa Nagamalli Santhoshkumar; Nan Niu; Upulee Kanewala

doi:10.1007/978-3-031-08760-8_42

Learning I/O Variables from Scientific Software’s User Manuals

Zedong Peng
, Xuanyi Lin
, Sreelekhaa Nagamalli Santhoshkumar
, Nan Niu
, Upulee Kanewala

Computer Science

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

3 Scopus citations

Abstract

Scientific software often involves many input and output variables. Identifying these variables is important for such software engineering tasks as metamorphic testing. To reduce the manual work, we report in this paper our investigation of machine learning algorithms in classifying variables from software’s user manuals. We identify thirteen natural-language features, and use them to develop a multi-layer solution where the first layer distinguishes variables from non-variables and the second layer classifies the variables into input and output types. Our experimental results on three scientific software systems show that random forest and feedforward neural network can be used to best implement the first layer and second layer respectively.

Original language	English
Title of host publication	Computational Science - ICCS 2022, 22nd International Conference, Proceedings
Editors	Derek Groen, Clélia de Mulatier, Valeria V. Krzhizhanovskaya, Peter M.A. Sloot, Maciej Paszynski, Jack J. Dongarra
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	503-516
Number of pages	14
ISBN (Print)	9783031087592
DOIs	https://doi.org/10.1007/978-3-031-08760-8_42
State	Published - 2022
Event	22nd Annual International Conference on Computational Science, ICCS 2022 - London, United Kingdom Duration: Jun 21 2022 → Jun 23 2022

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	13353 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	22nd Annual International Conference on Computational Science, ICCS 2022
Country/Territory	United Kingdom
City	London
Period	06/21/22 → 06/23/22

Keywords

Classification
Machine learning
Scientific software
Software documentation
User manual

Access to Document

10.1007/978-3-031-08760-8_42

Cite this

Peng, Z., Lin, X., Santhoshkumar, S. N., Niu, N., & Kanewala, U. (2022). Learning I/O Variables from Scientific Software’s User Manuals. In D. Groen, C. de Mulatier, V. V. Krzhizhanovskaya, P. M. A. Sloot, M. Paszynski, & J. J. Dongarra (Eds.), Computational Science - ICCS 2022, 22nd International Conference, Proceedings (pp. 503-516). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 13353 LNCS). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-031-08760-8_42

Peng, Zedong ; Lin, Xuanyi ; Santhoshkumar, Sreelekhaa Nagamalli et al. / Learning I/O Variables from Scientific Software’s User Manuals. Computational Science - ICCS 2022, 22nd International Conference, Proceedings. editor / Derek Groen ; Clélia de Mulatier ; Valeria V. Krzhizhanovskaya ; Peter M.A. Sloot ; Maciej Paszynski ; Jack J. Dongarra. Springer Science and Business Media Deutschland GmbH, 2022. pp. 503-516 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{bf40211b60d04f139450173c54d1f032,

title = "Learning I/O Variables from Scientific Software{\textquoteright}s User Manuals",

abstract = "Scientific software often involves many input and output variables. Identifying these variables is important for such software engineering tasks as metamorphic testing. To reduce the manual work, we report in this paper our investigation of machine learning algorithms in classifying variables from software{\textquoteright}s user manuals. We identify thirteen natural-language features, and use them to develop a multi-layer solution where the first layer distinguishes variables from non-variables and the second layer classifies the variables into input and output types. Our experimental results on three scientific software systems show that random forest and feedforward neural network can be used to best implement the first layer and second layer respectively.",

keywords = "Classification, Machine learning, Scientific software, Software documentation, User manual",

author = "Zedong Peng and Xuanyi Lin and Santhoshkumar, \{Sreelekhaa Nagamalli\} and Nan Niu and Upulee Kanewala",

note = "Publisher Copyright: {\textcopyright} 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.; 22nd Annual International Conference on Computational Science, ICCS 2022 ; Conference date: 21-06-2022 Through 23-06-2022",

year = "2022",

doi = "10.1007/978-3-031-08760-8\_42",

language = "English",

isbn = "9783031087592",

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

publisher = "Springer Science and Business Media Deutschland GmbH",

pages = "503--516",

editor = "Derek Groen and \{de Mulatier\}, Cl{\'e}lia and Krzhizhanovskaya, \{Valeria V.\} and Sloot, \{Peter M.A.\} and Maciej Paszynski and Dongarra, \{Jack J.\}",

booktitle = "Computational Science - ICCS 2022, 22nd International Conference, Proceedings",

}

Peng, Z, Lin, X, Santhoshkumar, SN, Niu, N & Kanewala, U 2022, Learning I/O Variables from Scientific Software’s User Manuals. in D Groen, C de Mulatier, VV Krzhizhanovskaya, PMA Sloot, M Paszynski & JJ Dongarra (eds), Computational Science - ICCS 2022, 22nd International Conference, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 13353 LNCS, Springer Science and Business Media Deutschland GmbH, pp. 503-516, 22nd Annual International Conference on Computational Science, ICCS 2022, London, United Kingdom, 06/21/22. https://doi.org/10.1007/978-3-031-08760-8_42

Learning I/O Variables from Scientific Software’s User Manuals. / Peng, Zedong; Lin, Xuanyi; Santhoshkumar, Sreelekhaa Nagamalli et al.
Computational Science - ICCS 2022, 22nd International Conference, Proceedings. ed. / Derek Groen; Clélia de Mulatier; Valeria V. Krzhizhanovskaya; Peter M.A. Sloot; Maciej Paszynski; Jack J. Dongarra. Springer Science and Business Media Deutschland GmbH, 2022. p. 503-516 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 13353 LNCS).

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

TY - GEN

T1 - Learning I/O Variables from Scientific Software’s User Manuals

AU - Peng, Zedong

AU - Lin, Xuanyi

AU - Santhoshkumar, Sreelekhaa Nagamalli

AU - Niu, Nan

AU - Kanewala, Upulee

PY - 2022

Y1 - 2022

N2 - Scientific software often involves many input and output variables. Identifying these variables is important for such software engineering tasks as metamorphic testing. To reduce the manual work, we report in this paper our investigation of machine learning algorithms in classifying variables from software’s user manuals. We identify thirteen natural-language features, and use them to develop a multi-layer solution where the first layer distinguishes variables from non-variables and the second layer classifies the variables into input and output types. Our experimental results on three scientific software systems show that random forest and feedforward neural network can be used to best implement the first layer and second layer respectively.

AB - Scientific software often involves many input and output variables. Identifying these variables is important for such software engineering tasks as metamorphic testing. To reduce the manual work, we report in this paper our investigation of machine learning algorithms in classifying variables from software’s user manuals. We identify thirteen natural-language features, and use them to develop a multi-layer solution where the first layer distinguishes variables from non-variables and the second layer classifies the variables into input and output types. Our experimental results on three scientific software systems show that random forest and feedforward neural network can be used to best implement the first layer and second layer respectively.

KW - Classification

KW - Machine learning

KW - Scientific software

KW - Software documentation

KW - User manual

UR - https://www.scopus.com/pages/publications/85134291806

U2 - 10.1007/978-3-031-08760-8_42

DO - 10.1007/978-3-031-08760-8_42

M3 - Conference contribution

AN - SCOPUS:85134291806

SN - 9783031087592

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

SP - 503

EP - 516

BT - Computational Science - ICCS 2022, 22nd International Conference, Proceedings

A2 - Groen, Derek

A2 - de Mulatier, Clélia

A2 - Krzhizhanovskaya, Valeria V.

A2 - Sloot, Peter M.A.

A2 - Paszynski, Maciej

A2 - Dongarra, Jack J.

PB - Springer Science and Business Media Deutschland GmbH

T2 - 22nd Annual International Conference on Computational Science, ICCS 2022

Y2 - 21 June 2022 through 23 June 2022

ER -

Peng Z, Lin X, Santhoshkumar SN, Niu N, Kanewala U. Learning I/O Variables from Scientific Software’s User Manuals. In Groen D, de Mulatier C, Krzhizhanovskaya VV, Sloot PMA, Paszynski M, Dongarra JJ, editors, Computational Science - ICCS 2022, 22nd International Conference, Proceedings. Springer Science and Business Media Deutschland GmbH. 2022. p. 503-516. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-031-08760-8_42