A Systematic Comparison of Search Algorithms for Topic Modelling—A Study on Duplicate Bug Report Identification

Annibale Panichella

doi:10.1007/978-3-030-27455-9_2

A Systematic Comparison of Search Algorithms for Topic Modelling—A Study on Duplicate Bug Report Identification

^*Corresponding author for this work

Software Engineering

Research output: Chapter in Book/Conference proceedings/Edited volume › Conference contribution › Scientific › peer-review

13 Citations (Scopus)

195 Downloads (Pure)

Abstract

Latent Dirichlet Allocation (LDA) has been used to support many software engineering tasks. Previous studies showed that default settings lead to sub-optimal topic modeling with a dramatic impact on the performance of such approaches in terms of precision and recall. For this reason, researchers used search algorithms (e.g., genetic algorithms) to automatically configure topic models in an unsupervised fashion. While previous work showed the ability of individual search algorithms in finding near-optimal configurations, it is not clear to what extent the choice of the meta-heuristic matters for SE tasks. In this paper, we present a systematic comparison of five different meta-heuristics to configure LDA in the context of duplicate bug reports identification. The results show that (1) no master algorithm outperforms the others for all software projects, (2) random search and PSO are the least effective meta-heuristics. Finally, the running time strongly depends on the computational complexity of LDA while the internal complexity of the search algorithms plays a negligible role.

Original language	English
Title of host publication	Search-Based Software Engineering - 11th International Symposium, SSBSE 2019, Proceedings
Editors	Shiva Nejati, Gregory Gay
Publisher	Springer
Pages	11-26
Number of pages	16
ISBN (Print)	9783030274542
DOIs	https://doi.org/10.1007/978-3-030-27455-9_2
Publication status	Published - 1 Jan 2019
Event	11th International Symposium on Search-Based Software Engineering, SSBSE 2019 - Tallinn, Estonia Duration: 31 Aug 2019 → 1 Sept 2019

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	11664 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	11th International Symposium on Search-Based Software Engineering, SSBSE 2019
Country/Territory	Estonia
City	Tallinn
Period	31/08/19 → 1/09/19

Keywords

Duplicate Bug Report
Evolutionary Algorithms
Latent Dirichlet Allocation
Search-based Software Engineering
Topic modeling

Access to Document

10.1007/978-3-030-27455-9_2

mainAccepted author manuscript, 379 KB

Cite this

Panichella, A. (2019). A Systematic Comparison of Search Algorithms for Topic Modelling—A Study on Duplicate Bug Report Identification. In S. Nejati, & G. Gay (Eds.), Search-Based Software Engineering - 11th International Symposium, SSBSE 2019, Proceedings (pp. 11-26). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11664 LNCS). Springer. https://doi.org/10.1007/978-3-030-27455-9_2

Panichella, Annibale. / A Systematic Comparison of Search Algorithms for Topic Modelling—A Study on Duplicate Bug Report Identification. Search-Based Software Engineering - 11th International Symposium, SSBSE 2019, Proceedings. editor / Shiva Nejati ; Gregory Gay. Springer, 2019. pp. 11-26 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{a3603e0f8d874735987650cc7aebf1f8,

title = "A Systematic Comparison of Search Algorithms for Topic Modelling—A Study on Duplicate Bug Report Identification",

abstract = "Latent Dirichlet Allocation (LDA) has been used to support many software engineering tasks. Previous studies showed that default settings lead to sub-optimal topic modeling with a dramatic impact on the performance of such approaches in terms of precision and recall. For this reason, researchers used search algorithms (e.g., genetic algorithms) to automatically configure topic models in an unsupervised fashion. While previous work showed the ability of individual search algorithms in finding near-optimal configurations, it is not clear to what extent the choice of the meta-heuristic matters for SE tasks. In this paper, we present a systematic comparison of five different meta-heuristics to configure LDA in the context of duplicate bug reports identification. The results show that (1) no master algorithm outperforms the others for all software projects, (2) random search and PSO are the least effective meta-heuristics. Finally, the running time strongly depends on the computational complexity of LDA while the internal complexity of the search algorithms plays a negligible role.",

keywords = "Duplicate Bug Report, Evolutionary Algorithms, Latent Dirichlet Allocation, Search-based Software Engineering, Topic modeling",

author = "Annibale Panichella",

year = "2019",

month = jan,

day = "1",

doi = "10.1007/978-3-030-27455-9_2",

language = "English",

isbn = "9783030274542",

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

publisher = "Springer",

pages = "11--26",

editor = "Shiva Nejati and Gregory Gay",

booktitle = "Search-Based Software Engineering - 11th International Symposium, SSBSE 2019, Proceedings",

note = "11th International Symposium on Search-Based Software Engineering, SSBSE 2019 ; Conference date: 31-08-2019 Through 01-09-2019",

}

Panichella, A 2019, A Systematic Comparison of Search Algorithms for Topic Modelling—A Study on Duplicate Bug Report Identification. in S Nejati & G Gay (eds), Search-Based Software Engineering - 11th International Symposium, SSBSE 2019, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 11664 LNCS, Springer, pp. 11-26, 11th International Symposium on Search-Based Software Engineering, SSBSE 2019, Tallinn, Estonia, 31/08/19. https://doi.org/10.1007/978-3-030-27455-9_2

A Systematic Comparison of Search Algorithms for Topic Modelling—A Study on Duplicate Bug Report Identification. / Panichella, Annibale.
Search-Based Software Engineering - 11th International Symposium, SSBSE 2019, Proceedings. ed. / Shiva Nejati; Gregory Gay. Springer, 2019. p. 11-26 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11664 LNCS).

Research output: Chapter in Book/Conference proceedings/Edited volume › Conference contribution › Scientific › peer-review

TY - GEN

T1 - A Systematic Comparison of Search Algorithms for Topic Modelling—A Study on Duplicate Bug Report Identification

AU - Panichella, Annibale

PY - 2019/1/1

Y1 - 2019/1/1

N2 - Latent Dirichlet Allocation (LDA) has been used to support many software engineering tasks. Previous studies showed that default settings lead to sub-optimal topic modeling with a dramatic impact on the performance of such approaches in terms of precision and recall. For this reason, researchers used search algorithms (e.g., genetic algorithms) to automatically configure topic models in an unsupervised fashion. While previous work showed the ability of individual search algorithms in finding near-optimal configurations, it is not clear to what extent the choice of the meta-heuristic matters for SE tasks. In this paper, we present a systematic comparison of five different meta-heuristics to configure LDA in the context of duplicate bug reports identification. The results show that (1) no master algorithm outperforms the others for all software projects, (2) random search and PSO are the least effective meta-heuristics. Finally, the running time strongly depends on the computational complexity of LDA while the internal complexity of the search algorithms plays a negligible role.

AB - Latent Dirichlet Allocation (LDA) has been used to support many software engineering tasks. Previous studies showed that default settings lead to sub-optimal topic modeling with a dramatic impact on the performance of such approaches in terms of precision and recall. For this reason, researchers used search algorithms (e.g., genetic algorithms) to automatically configure topic models in an unsupervised fashion. While previous work showed the ability of individual search algorithms in finding near-optimal configurations, it is not clear to what extent the choice of the meta-heuristic matters for SE tasks. In this paper, we present a systematic comparison of five different meta-heuristics to configure LDA in the context of duplicate bug reports identification. The results show that (1) no master algorithm outperforms the others for all software projects, (2) random search and PSO are the least effective meta-heuristics. Finally, the running time strongly depends on the computational complexity of LDA while the internal complexity of the search algorithms plays a negligible role.

KW - Duplicate Bug Report

KW - Evolutionary Algorithms

KW - Latent Dirichlet Allocation

KW - Search-based Software Engineering

KW - Topic modeling

UR - http://www.scopus.com/inward/record.url?scp=85072853501&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-27455-9_2

DO - 10.1007/978-3-030-27455-9_2

M3 - Conference contribution

AN - SCOPUS:85072853501

SN - 9783030274542

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

SP - 11

EP - 26

BT - Search-Based Software Engineering - 11th International Symposium, SSBSE 2019, Proceedings

A2 - Nejati, Shiva

A2 - Gay, Gregory

PB - Springer

T2 - 11th International Symposium on Search-Based Software Engineering, SSBSE 2019

Y2 - 31 August 2019 through 1 September 2019

ER -

Panichella A. A Systematic Comparison of Search Algorithms for Topic Modelling—A Study on Duplicate Bug Report Identification. In Nejati S, Gay G, editors, Search-Based Software Engineering - 11th International Symposium, SSBSE 2019, Proceedings. Springer. 2019. p. 11-26. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-030-27455-9_2

A Systematic Comparison of Search Algorithms for Topic Modelling—A Study on Duplicate Bug Report Identification

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

SSBSE 2019 Best Paper Award

Cite this

A Systematic Comparison of Search Algorithms for Topic Modelling—A Study on Duplicate Bug Report Identification

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Prizes

SSBSE 2019 Best Paper Award

Cite this