Classification of speaking proficiency level by machine learning and feature selection

Brendan Flanagan; Sachio Hirokawa; Emiko Kaneko; Emi Izumi

doi:10.1007/978-3-319-52836-6_72

Classification of speaking proficiency level by machine learning and feature selection

Brendan Flanagan, Sachio Hirokawa, Emiko Kaneko, Emi Izumi

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

Abstract

Analysis of publicly available language learning corpora can be useful for extracting characteristic features of learners from different proficiency levels. This can then be used to support language learning research and the creation of educational resources. In this paper, we classify the words and parts of speech of transcripts from different speaking proficiency levels found in the NICT-JLE corpus. The characteristic features of learners who have the equivalent spoken proficiency of CEFR levels A1 through to B2 were extracted by analyzing the data with the support vector machine method. In particular, we apply feature selection to find a set of characteristic features that achieve optimal classification performance, which can be used to predict spoken learner proficiency.

Original language	English
Title of host publication	Emerging Technologies for Education - 1st International Symposium, SETE 2016 Held in Conjunction with ICWL 2016, Revised Selected Papers
Editors	Rosella Gennari, Yiwei Cao, Yueh-Min Huang, Wu Wu, Haoran Xie
Publisher	Springer Verlag
Pages	677-682
Number of pages	6
ISBN (Print)	9783319528359
DOIs	https://doi.org/10.1007/978-3-319-52836-6_72
Publication status	Published - Jan 1 2017
Event	1st International Symposium on Emerging Technologies for Education, SETE 2016 Held in Conjunction with ICWL 2016 - Rome, Italy Duration: Oct 26 2016 → Oct 29 2016

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	10108 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Other

Other	1st International Symposium on Emerging Technologies for Education, SETE 2016 Held in Conjunction with ICWL 2016
Country	Italy
City	Rome
Period	10/26/16 → 10/29/16

All Science Journal Classification (ASJC) codes

Theoretical Computer Science
Computer Science(all)

Access to Document

10.1007/978-3-319-52836-6_72

Fingerprint Dive into the research topics of 'Classification of speaking proficiency level by machine learning and feature selection'. Together they form a unique fingerprint.

Cite this

Flanagan, B., Hirokawa, S., Kaneko, E., & Izumi, E. (2017). Classification of speaking proficiency level by machine learning and feature selection. In R. Gennari, Y. Cao, Y-M. Huang, W. Wu, & H. Xie (Eds.), Emerging Technologies for Education - 1st International Symposium, SETE 2016 Held in Conjunction with ICWL 2016, Revised Selected Papers (pp. 677-682). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10108 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-319-52836-6_72

Classification of speaking proficiency level by machine learning and feature selection. / Flanagan, Brendan; Hirokawa, Sachio; Kaneko, Emiko; Izumi, Emi.

Emerging Technologies for Education - 1st International Symposium, SETE 2016 Held in Conjunction with ICWL 2016, Revised Selected Papers. ed. / Rosella Gennari; Yiwei Cao; Yueh-Min Huang; Wu Wu; Haoran Xie. Springer Verlag, 2017. p. 677-682 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10108 LNCS).

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

Flanagan, B, Hirokawa, S, Kaneko, E & Izumi, E 2017, Classification of speaking proficiency level by machine learning and feature selection. in R Gennari, Y Cao, Y-M Huang, W Wu & H Xie (eds), Emerging Technologies for Education - 1st International Symposium, SETE 2016 Held in Conjunction with ICWL 2016, Revised Selected Papers. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 10108 LNCS, Springer Verlag, pp. 677-682, 1st International Symposium on Emerging Technologies for Education, SETE 2016 Held in Conjunction with ICWL 2016, Rome, Italy, 10/26/16. https://doi.org/10.1007/978-3-319-52836-6_72

Flanagan B, Hirokawa S, Kaneko E, Izumi E. Classification of speaking proficiency level by machine learning and feature selection. In Gennari R, Cao Y, Huang Y-M, Wu W, Xie H, editors, Emerging Technologies for Education - 1st International Symposium, SETE 2016 Held in Conjunction with ICWL 2016, Revised Selected Papers. Springer Verlag. 2017. p. 677-682. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-319-52836-6_72

Flanagan, Brendan ; Hirokawa, Sachio ; Kaneko, Emiko ; Izumi, Emi. / Classification of speaking proficiency level by machine learning and feature selection. Emerging Technologies for Education - 1st International Symposium, SETE 2016 Held in Conjunction with ICWL 2016, Revised Selected Papers. editor / Rosella Gennari ; Yiwei Cao ; Yueh-Min Huang ; Wu Wu ; Haoran Xie. Springer Verlag, 2017. pp. 677-682 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{5e4112b7d4e94b73a64c097fd9f7a91d,

title = "Classification of speaking proficiency level by machine learning and feature selection",

abstract = "Analysis of publicly available language learning corpora can be useful for extracting characteristic features of learners from different proficiency levels. This can then be used to support language learning research and the creation of educational resources. In this paper, we classify the words and parts of speech of transcripts from different speaking proficiency levels found in the NICT-JLE corpus. The characteristic features of learners who have the equivalent spoken proficiency of CEFR levels A1 through to B2 were extracted by analyzing the data with the support vector machine method. In particular, we apply feature selection to find a set of characteristic features that achieve optimal classification performance, which can be used to predict spoken learner proficiency.",

author = "Brendan Flanagan and Sachio Hirokawa and Emiko Kaneko and Emi Izumi",

year = "2017",

month = jan,

day = "1",

doi = "10.1007/978-3-319-52836-6_72",

language = "English",

isbn = "9783319528359",

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

publisher = "Springer Verlag",

pages = "677--682",

editor = "Rosella Gennari and Yiwei Cao and Yueh-Min Huang and Wu Wu and Haoran Xie",

booktitle = "Emerging Technologies for Education - 1st International Symposium, SETE 2016 Held in Conjunction with ICWL 2016, Revised Selected Papers",

address = "Germany",

note = "1st International Symposium on Emerging Technologies for Education, SETE 2016 Held in Conjunction with ICWL 2016 ; Conference date: 26-10-2016 Through 29-10-2016",

}

TY - GEN

T1 - Classification of speaking proficiency level by machine learning and feature selection

AU - Flanagan, Brendan

AU - Hirokawa, Sachio

AU - Kaneko, Emiko

AU - Izumi, Emi

PY - 2017/1/1

Y1 - 2017/1/1

N2 - Analysis of publicly available language learning corpora can be useful for extracting characteristic features of learners from different proficiency levels. This can then be used to support language learning research and the creation of educational resources. In this paper, we classify the words and parts of speech of transcripts from different speaking proficiency levels found in the NICT-JLE corpus. The characteristic features of learners who have the equivalent spoken proficiency of CEFR levels A1 through to B2 were extracted by analyzing the data with the support vector machine method. In particular, we apply feature selection to find a set of characteristic features that achieve optimal classification performance, which can be used to predict spoken learner proficiency.

AB - Analysis of publicly available language learning corpora can be useful for extracting characteristic features of learners from different proficiency levels. This can then be used to support language learning research and the creation of educational resources. In this paper, we classify the words and parts of speech of transcripts from different speaking proficiency levels found in the NICT-JLE corpus. The characteristic features of learners who have the equivalent spoken proficiency of CEFR levels A1 through to B2 were extracted by analyzing the data with the support vector machine method. In particular, we apply feature selection to find a set of characteristic features that achieve optimal classification performance, which can be used to predict spoken learner proficiency.

UR - http://www.scopus.com/inward/record.url?scp=85014242435&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85014242435&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-52836-6_72

DO - 10.1007/978-3-319-52836-6_72

M3 - Conference contribution

AN - SCOPUS:85014242435

SN - 9783319528359

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

SP - 677

EP - 682

BT - Emerging Technologies for Education - 1st International Symposium, SETE 2016 Held in Conjunction with ICWL 2016, Revised Selected Papers

A2 - Gennari, Rosella

A2 - Cao, Yiwei

A2 - Huang, Yueh-Min

A2 - Wu, Wu

A2 - Xie, Haoran

PB - Springer Verlag

T2 - 1st International Symposium on Emerging Technologies for Education, SETE 2016 Held in Conjunction with ICWL 2016

Y2 - 26 October 2016 through 29 October 2016

ER -