Web-Based Concordancer to Learn Usage of English Expressions Takashi Yamanoue Kyushu Institute of Technology, Japan Toshiro Minami Kyushu Institute of.

Web-Based Concordancer to Learn Usage of English Expressions

Takashi Yamanoue Kyushu Institute of Technology, JapanToshiro Minami Kyushu Institute of Information Sciences &

Kyushu University, JapanIan Ruxton Kyushu Institute of Technology, Japan

ICITA2002@Bathurst (02.11.25-28)

Contents

Motivation System Design and Implementation Examples and Experiments for

Evaluation Related Work Concluding Remarks

Difficulties in writing in English

Is it really used??

?

•Spelling → spell checker•Grammar → grammar checker•Usage → ?

Motivation

Help for Usage

Corpus linguistics Time-consuming Needs hard work in order to make a good

corpus Copyright problem

(often) Outdated from the beginning Tools are mainly for experts

Difficult to use for ordinary learners

A Solution

Use of “Web-Corpus”= Using Web Documents as a Corpus

Maintenance free: Exists as it is Always new, reflects current status of

languagesA lot of applications/services are available on

the Internet

System Design and Implementation（ WebLEAP ）

Features

With WebLEAP(Web Language Evaluation Assistant Program) , we can get some information for: comparing expressions, deciding if the expression is right or wrong, and finding out wrong parts.

System Organization

WebLEAP

See/watch the TV/movie

“the TV” “see the TV” “watch the TV”

216,107 873* 1,593

“the movie” “see the movie” “watch the movie”

472,776 13,638 6,666*

common.

Not common

Examples and Experiments for Evaluation

ExperimentsPerson’s Name

Place Names

“Bertrand Russell” “Burtrand Russell”

12,575 3

Not only for Historical names

“Jenolan caves” “Genolan caves”

992 2

Choose one phrase or word that should be corrected.

Successful : 13/20 Insufficient information: 2/20Failure: 5/20

TOEIC(Test of English for International Communication)

Only 9.4% of the testees gave the correct answer

Successful Example:

Failure Example:

Comparison with others

WebLEAP detects 13(+2) /20Checker1 detects 7 /20Checker2 detects 4 /20WebLEAP shows not only places but also pop

ularity(degree).

WebLEAP Experimental Class11 studentsWriting essays for 30 min.

2 classes with same membersCompare the numbers of errors

Results

StudentEssay1 Essay2

errors words density errors words density

1 5 53 9% 4 48 8% ↓

2 9 43 21% 6 32 19% ↓

3 5 31 16% 4 36 11% ↓

4 4 34 12% 7 33 21%

5 5 34 15% 6 42 14% ↓

6 5 34 15% 4 39 10% ↓

7 5 39 13% 8 40 20%

8 6 54 11% 1 11 9% ↓

9 3 35 9% 3 40 8% ↓

10 4 34 12% 5 32 16%

11 4 33 12% 7 32 22%

Evaluation

More than half of the students got the decrease of the error densities.

(Current) WebLEAP is not good enough for beginners of English.

Thus for most of them are in the “TGR(Teachers Guidance Recommended)” level.

All the students enjoyed using WebLEAP. WebLEAP makes us think and interpret the expressio

ns we use. → Educational Effect

Satoh’s system… WWW, KWICWordSmith Tools…for expertsTXTANA … for expertsSARA…for research purposeSUIKO…detect wrong sentences The Writer’s Assistant…idea organizer

Related Work

Sato’s System: Input

Sato’s System: Results

WebLEAP: is a tool for helping with writing. gives us information To know, is it really used? is not for beginners.

Need skill to interpret the given results. is such a system that users are excited about us

ing it.

Concluding Remarks

Integrate with Web-based and other types of services

Monitor user’s behavior and get tips for improvement

Assist collaborative writing

Further Research Topics

Thank you for Listening!