No Cover Image

E-Thesis 101 views 32 downloads

Automatic essay scoring for low level learners of English as a second language. / Andrew Mellor

Swansea University Author: Andrew, Mellor

Abstract

This thesis investigates the automatic assessment of essays written by Japanese low level learners of English as a second language. A number of essay features are investigated for their ability to predict human assessments of quality. These features include unique lexical signatures (Meara. Jacobs &...

Full description

Published: 2010
Institution: Swansea University
Degree level: Doctoral
Degree name: Ph.D
URI: https://cronfa.swan.ac.uk/Record/cronfa42247
Tags: Add Tag
No Tags, Be the first to tag this record!
first_indexed 2018-08-02T18:54:14Z
last_indexed 2018-08-03T10:09:38Z
id cronfa42247
recordtype RisThesis
fullrecord <?xml version="1.0"?><rfc1807><datestamp>2018-08-02T16:24:28.5577850</datestamp><bib-version>v2</bib-version><id>42247</id><entry>2018-08-02</entry><title>Automatic essay scoring for low level learners of English as a second language.</title><swanseaauthors><author><sid>55856f6a55b7ac9ef7933c538a08b207</sid><ORCID>NULL</ORCID><firstname>Andrew</firstname><surname>Mellor</surname><name>Andrew Mellor</name><active>true</active><ethesisStudent>true</ethesisStudent></author></swanseaauthors><date>2018-08-02</date><abstract>This thesis investigates the automatic assessment of essays written by Japanese low level learners of English as a second language. A number of essay features are investigated for their ability to predict human assessments of quality. These features include unique lexical signatures (Meara. Jacobs &amp; Rodgers, 2002), distinctiveness, essay length, various measures of lexical diversity, mean sentence length and some properties of word distributions. Findings suggest that no one feature is sufficient to account for essay quality but essay length is a strong predictor for low level learners in time constrained tasks. Combinations of several features are much more powerful in predicting quality than single features. Some simple systems incorporating some of these features are also considered. One is a two-dimensional 'quantity/content' model based on essay length and lexical diversity. Various measures of lexical diversity are used for the content dimension. Another system considered is a clustering algorithm based on various lexical features. A third system is a Bayesian algorithm which classifies essays according to semantic content. Finally, an alternative process based on capture-recapture analysis is also considered for special cases of assessment. One interesting finding is that although many essay features only have moderate associations with quality, extreme values at both ends of the scale are often very reliable indicators of high quality' or poor quality essays. These easily identifiable high quality or low quality essays can act as training samples for classification algorithms such as Bayesian classifiers. The clustering algorithm used in this study correlated particularly strongly with human essay ratings. This suggests that multivariate statistical methods may help realise more accurate essay prediction.</abstract><type>E-Thesis</type><journal/><journalNumber></journalNumber><paginationStart/><paginationEnd/><publisher/><placeOfPublication/><isbnPrint/><issnPrint/><issnElectronic/><keywords>English as a second language.</keywords><publishedDay>31</publishedDay><publishedMonth>12</publishedMonth><publishedYear>2010</publishedYear><publishedDate>2010-12-31</publishedDate><doi/><url/><notes/><college>COLLEGE NANME</college><department>English Language and Applied Linguistics</department><CollegeCode>COLLEGE CODE</CollegeCode><institution>Swansea University</institution><degreelevel>Doctoral</degreelevel><degreename>Ph.D</degreename><apcterm/><lastEdited>2018-08-02T16:24:28.5577850</lastEdited><Created>2018-08-02T16:24:28.5577850</Created><path><level id="1">College of Arts and Humanities</level><level id="2">English Language and Applied Linguistics</level></path><authors><author><firstname>Andrew</firstname><surname>Mellor</surname><orcid>NULL</orcid><order>1</order></author></authors><documents><document><filename>0042247-02082018162439.pdf</filename><originalFilename>10797955.pdf</originalFilename><uploaded>2018-08-02T16:24:39.5570000</uploaded><type>Output</type><contentLength>7317685</contentLength><contentType>application/pdf</contentType><version>E-Thesis</version><cronfaStatus>true</cronfaStatus><action/><embargoDate>2018-08-02T16:24:39.5570000</embargoDate><copyrightCorrect>false</copyrightCorrect></document></documents><OutputDurs/></rfc1807>
spelling 2018-08-02T16:24:28.5577850 v2 42247 2018-08-02 Automatic essay scoring for low level learners of English as a second language. 55856f6a55b7ac9ef7933c538a08b207 NULL Andrew Mellor Andrew Mellor true true 2018-08-02 This thesis investigates the automatic assessment of essays written by Japanese low level learners of English as a second language. A number of essay features are investigated for their ability to predict human assessments of quality. These features include unique lexical signatures (Meara. Jacobs & Rodgers, 2002), distinctiveness, essay length, various measures of lexical diversity, mean sentence length and some properties of word distributions. Findings suggest that no one feature is sufficient to account for essay quality but essay length is a strong predictor for low level learners in time constrained tasks. Combinations of several features are much more powerful in predicting quality than single features. Some simple systems incorporating some of these features are also considered. One is a two-dimensional 'quantity/content' model based on essay length and lexical diversity. Various measures of lexical diversity are used for the content dimension. Another system considered is a clustering algorithm based on various lexical features. A third system is a Bayesian algorithm which classifies essays according to semantic content. Finally, an alternative process based on capture-recapture analysis is also considered for special cases of assessment. One interesting finding is that although many essay features only have moderate associations with quality, extreme values at both ends of the scale are often very reliable indicators of high quality' or poor quality essays. These easily identifiable high quality or low quality essays can act as training samples for classification algorithms such as Bayesian classifiers. The clustering algorithm used in this study correlated particularly strongly with human essay ratings. This suggests that multivariate statistical methods may help realise more accurate essay prediction. E-Thesis English as a second language. 31 12 2010 2010-12-31 COLLEGE NANME English Language and Applied Linguistics COLLEGE CODE Swansea University Doctoral Ph.D 2018-08-02T16:24:28.5577850 2018-08-02T16:24:28.5577850 College of Arts and Humanities English Language and Applied Linguistics Andrew Mellor NULL 1 0042247-02082018162439.pdf 10797955.pdf 2018-08-02T16:24:39.5570000 Output 7317685 application/pdf E-Thesis true 2018-08-02T16:24:39.5570000 false
title Automatic essay scoring for low level learners of English as a second language.
spellingShingle Automatic essay scoring for low level learners of English as a second language.
Andrew, Mellor
title_short Automatic essay scoring for low level learners of English as a second language.
title_full Automatic essay scoring for low level learners of English as a second language.
title_fullStr Automatic essay scoring for low level learners of English as a second language.
title_full_unstemmed Automatic essay scoring for low level learners of English as a second language.
title_sort Automatic essay scoring for low level learners of English as a second language.
author_id_str_mv 55856f6a55b7ac9ef7933c538a08b207
author_id_fullname_str_mv 55856f6a55b7ac9ef7933c538a08b207_***_Andrew, Mellor
author Andrew, Mellor
author2 Andrew Mellor
format E-Thesis
publishDate 2010
institution Swansea University
college_str College of Arts and Humanities
hierarchytype
hierarchy_top_id collegeofartsandhumanities
hierarchy_top_title College of Arts and Humanities
hierarchy_parent_id collegeofartsandhumanities
hierarchy_parent_title College of Arts and Humanities
department_str English Language and Applied Linguistics{{{_:::_}}}College of Arts and Humanities{{{_:::_}}}English Language and Applied Linguistics
document_store_str 1
active_str 0
description This thesis investigates the automatic assessment of essays written by Japanese low level learners of English as a second language. A number of essay features are investigated for their ability to predict human assessments of quality. These features include unique lexical signatures (Meara. Jacobs & Rodgers, 2002), distinctiveness, essay length, various measures of lexical diversity, mean sentence length and some properties of word distributions. Findings suggest that no one feature is sufficient to account for essay quality but essay length is a strong predictor for low level learners in time constrained tasks. Combinations of several features are much more powerful in predicting quality than single features. Some simple systems incorporating some of these features are also considered. One is a two-dimensional 'quantity/content' model based on essay length and lexical diversity. Various measures of lexical diversity are used for the content dimension. Another system considered is a clustering algorithm based on various lexical features. A third system is a Bayesian algorithm which classifies essays according to semantic content. Finally, an alternative process based on capture-recapture analysis is also considered for special cases of assessment. One interesting finding is that although many essay features only have moderate associations with quality, extreme values at both ends of the scale are often very reliable indicators of high quality' or poor quality essays. These easily identifiable high quality or low quality essays can act as training samples for classification algorithms such as Bayesian classifiers. The clustering algorithm used in this study correlated particularly strongly with human essay ratings. This suggests that multivariate statistical methods may help realise more accurate essay prediction.
published_date 2010-12-31T03:58:26Z
_version_ 1706948670881005568
score 10.810468