This book is the revised version of the author's Ph.D. dissertation which has proposed and implemented the combined model of grammar checking. It has reviewed in the first half of the book some related technologies to the present research, including POS tagging,parsing, grammar checking and automated essay scoring. Heart-felt gratitude shall be extended to the authors and reviewers of these researches. It is they that have helped me better understand the state-of-the-art of these technologies and have laid a solid foundation for the present research. In the second half of the book the author has described the proposed combined model and evaluated the application of the present proposal in computer-assisted semiautomated essay scoring in TESOL setting in China.

Preface

List of Abbreviations

Chapter I Introduction

 1.1 Topics Covered in the Book

 1.2 Outline of This Book

Chapter 2 From Tagging to Scoring

 2.1 Introduction

 2.2 Starting from Tagging and Parsing

2.2.1 Tagging

2.2.2 Parsing

 2.3 Grammar Checking

2.3.1 Researches Review

2.3.2 Projects Review

 2.4 Automated Essay Scoring

2.4.1 Project Essay Grade (PEG)

2.4.2 Latent Semantic Analysis (LSA)

2.4.3 Electronic Essay Rater (E-rater)

2.4.4 Bayesian Essay Test Scoring System (BETSY)

 2.5 Summary

Chapter3 Dependency Grammars

 3.1 Introduction

 3.2 Dependency Grammar

3.2.1 Lucien Tesni6re

3.2.2 Post-Tesni6re Development

3.2.3 Phrase Structure Grammars vs. Dependency Grammars

3.2.4 Link Grammar

 3.3 Link Grammar: A Detailed Discussion

3.3.1 Links

3.3.2 Basic Assumptions

3.3.3 Dictionary and Linking Requirements

3.3.4 Post-processing

3.3.5 Speed and Robustness Features

3.3.6 The Cost System

3.3.7 Link Grammar vs. Dependency-related Grafnmars

 3.4 Summary

Chapter 4 TBL Learning

 4.1 Introduction

 4.2 Tagging

 4.3 Parsing

 4.4 The Comparison Algorithm

 4.5 Summary

Chapter 5 Combined Model of Grammar Checking

 5.1 Introduction

 5.2 Combined Modeling

 5.3 Link Parser Revisited

5.3.1 Suitability for the Purpose of Grammar Checking

5.3.2 Evaluation of the Parser

5.3.3 A closer examination

5.3.4 Problems to be Addressed

 5.4 Phrase Structure Information Incorporated

 5.5 Summary

Chapter 6 Algorithm Implementation

 6.1 Research Questions

 6.2 Objectives of This Research

 6.3 Research Methodology

 6.4 Research Design

6.4.1 Prepanngthe Training Corpus

6.4.2 Training in TBL Algorithm

6.4.3 Applying the Knowledge to the Post Processing Stage in Link Parser

6.4.4 Evaluating the Combined Model

 6.5 Grammar Checker

6.5.1 The Interface

6.5.2 Speed of the Grammar Checker

6.5.3 Application of the Grammar Checker

6.5.4 Evaluation of the Grammar Checker

 6.6 Summary

Chapter 7 Evaluating the Combined Model

 7.1 Introduction

 7.2 Experiment Ⅰ

7.2.1 Test Set

7.2.2 Individual Sentence Observation

7.2.3 Overall Observation

 7.3 Experiment Ⅱ

7.3.1 Test Set

7.3.2 Results and Analysis

7.3.3 Case Study

7.3.4 Linkage-effect Relations

 7.4 Summary

Chapter 8 Evaluating the Grammar Checker

 8.1 Performance of the Grammar Checker

 8.2 Experiment III: A Comprehensive Test

 8.3 Summary

Chapter 9 Application in Automated Essay Scoring

 9.1 Introduction

 9.2 Training of the Grammar Checker Module

9.2.1 Training Set

9.2.2 Preliminary Regression

9.2.3 Factor Analysis

9.2.4 Final Regression

 9.3 Testing the Grammar Checker Module

9.3.1 Accuracy Test

9.3.2 Discussion

 9.4 Summary

Chapter 10 Remaining Issues and Future Work

 10.1 Error and Its Judgment

 10.2 Adaptiveness on CLEC

 10.3 Code-level Refinement

Appendix A Comparison Algorithm (in Perl Code)

Appendix B The Rule Set Learned

Appendix C Re-ranking in Experiment Ⅱ

Bibliography

Index