Computational Linguistics, Speech And Image Processing For Arabic Language

by Neamat El Gayar, Ching Yee Suen

Overview

This book encompasses a collection of topics covering recent advances that are important to the Arabic language in areas of natural language processing, speech and image analysis. This book presents state-of-the-art reviews and fundamentals as well as applications and recent innovations.The book chapters by top researchers present basic concepts and challenges for the Arabic language in linguistic processing, handwritten recognition, document analysis, text classification and speech processing. In addition, it reports on selected applications in sentiment analysis, annotation, text summarization, speech and font analysis, word recognition and spotting and question answering.Moreover, it highlights and introduces some novel applications in vital areas for the Arabic language. The book is therefore a useful resource for young researchers who are interested in the Arabic language and are still developing their fundamentals and skills in this area. It is also interesting for scientists who wish to keep track of the most recent research directions and advances in this area.

Product Details

ISBN-13:	9789813229389
Publisher:	World Scientific Publishing Company, Incorporated
Publication date:	11/20/2018
Series:	Series On Language Processing, Pattern Recognition, And Intelligent Systems , #4
Pages:	288
Product dimensions:	6.00(w) x 9.00(h) x 0.69(d)

Table of Contents

Preface v

Chapter 1 Arabic Speech Recognition: Challenges and State of the Art Sherif Mahdy Abdou Abdullah M. Moussa 1

1 Introduction 2

2 The Automatic Speech Recognition System Components 2

2.1 Pronunciation lexicon 4

2.2 Acoustic model 4

2.3 Language model 8

2.4 Decoding 9

3 Literature Review for Arabic ASR 10

4 Challenges for Arabic ASR Systems 14

4.1 Using non-diacritized Arabic data 15

4.2 Speech recognition for Arabic dialects 16

4.3 Inflection effect and the large vocabulary 19

5 State of the Art Arabic ASR Performance 22

6 Conclusions 24

References 24

Chapter 2 Introduction to Arabic Computational Linguistics Mohsen Rashwan 29

1 Introduction 29

2 Layers of Linguistic Analysis 30

2.1 Phonological analysis 30

2.2 Morphological analysis 31

2.3 Syntactic analysis 31

2.4 Semantic analysis 31

3 Challenges Facing Human Language Technologies 32

4 Challenges Facing the Arabic Language Processing 32

4.1 Arabic script 33

4.2 Common mistakes 33

4.3 Morphological structure for the Arabic word 34

4.4 Syntax of the Arabic sentence 35

5 Defining the Human Languages Technologies 36

5.1 Texts search (search engines) 36

5.2 Machine translation 38

5.3 Question answering 39

5.4 Automated essay scoring 39

5.5 Automatic text summarization 40

5.6 Document classification and clustering 40

5.7 Opinion mining 41

5.8 Computer-aided language learning (CALL) 42

5.9 Stylometry 42

5.10 Automatic speech recognition 43

5.11 Text to speech (TTS) 45

5.12 Audio and video search 46

5.13 Language recognition 46

5.14 Computer-aided pronunciation learning 46

5.15 Typewritten optical character recognition (OCR) 47

5.16 Intelligent character recognition 48

5.17 Book reader 48

5.18 Speech to speech translation 49

5.19 Speech-to-sign-language and sign-language-to-speech 49

5.20 Dialog management systems 50

5.21 Advanced information retrieval systems 51

5.22 Text mining (TM) 52

6 Arabic Computational Linguistics Institutions 52

6.1 Academic institutions 52

6.2 Companies interested in computational linguistics 56

7 Summary and Conclusions 57

References 57

Chapter 3 Challenges in Arabic Natural Language Processing Khaled Shaalan Sanjeera Siddiqui Manar Alkhatib Azza Abdel Monem 59

1 Introduction 59

2 Challenges 61

2.1 Arabic orthography 62

2.2 Arabic morphology 69

2.3 Syntax is intricate 72

3 Conclusion 78

References 79

Chapter 4 Arabic Recognition Based on Statistical Methods A. Belaïd A. Kacem Echi 85

1 Introduction 85

2 A Challenging Morphology 86

3 Features Extraction Techniques 87

4 Machine Learning Techniques 92

5 Markov Models 94

5.1 Case 1: Decomposition of the shape/label 94

5.2 Case 2: Decomposition by association with a model 96

5.3 Extension of HMM to the Plane 98

5.4 Bayesian Networks 99

5.5 Two Dimensional HMM 101

6 Discriminative Models 103

7 Conclusion 107

References 108

Chapter 5 Arabic Word Spotting Approaches and Techniques Muna Khayyat Louisa Lam Ching Y. Suen 111

1 Word Spotting 111

1.1 Definition 112

1.2 Input queries 113

1.3 Performance measures 114

1.4 Word spotting approaches 115

2 Arabic Word Spotting 116

2.1 Characteristics of Arabic handwriting 116

2.2 Arabic word spotting approaches 118

3 Databases 120

4 Extracted Features 121

5 Concluding Remarks 123

References 123

Chapter 6 A'rib - A Tool to Facilitate School Children's Ability to Analyze Arabic Sentences Syntactically Mashael Almedlej Aqil M Azmi 127

1 Introduction 127

2 Related Work 130

3 Basic Arabic Sentences Structure 131

4 System Design 132

4.1 Lexical analyzer 134

4.2 Syntactic analyzer 134

4.3 Results builder 138

4.4 Special cases 139

5 Implementation 140

5.1 Lexical analysis 141

5.2 Syntactic analysis 145

5.3 Results builder 151

5.4 Output 152

6 Conclusion and Future Work 152

References 153

Chapter 7 Semi-Automatic Data Annotation, POS Tagging and Mildly Context-Sensitive Disambiguation: The extended Revised AraMorph (XRAM) Giuliano Lancioni Laura Garofalo Raoul Villano Francesca Romana Romani Marta Campanelli Ilaria Cicola Ivana Pepe Valeria Pettinari Simona Olivieri 155

1 Introduction 155

2 Description of XRAM 156

2.1 Flag-selectable usage markers 157

2.2 Probabilistic mildly context-sensitive annotation 160

2.3 Lexical and morphological XML tagging of texts 161

2.4 Semi-automatic increment of lexical coverage 163

3 Validation and Research Grounds 165

4 Conclusion 166

References 166

Chapter 8 WeightedNileULex: A Scored Arabic Sentiment Lexicon for Improved Sentiment Analysis Samhaa R. El-Beltagy 169

1 Introduction 169

2 Related Work 170

3 The Base Lexicon 172

4 Assigning Scores to Lexicon Entries 173

4.1 Data collection 173

4.2 Collecting term statistics 174

4.3 Term scoring 174

5 Experiments and Results 178

5.1 The sentiment analysis system 179

5.2 The used datasets 180

5.3 Experimental results 181

6 Conclusion 184

References 184

Chapter 9 Islamic Fatwa Request Routing via Hierarchical Multi-Label Arabic Text Categorization? Reda Zayed Mohamed Farouk Hesham Hefny 187

1 Introduction 187

2 Related Work 190

3 Islamic Fatwa Requests Routing System 191

3.1 Text preprocessing 191

3.2 Feature engineering 193

3.3 The HOMER algorithm 194

4 Performance Evaluation 195

4.1 Data description 195

4.2 Methods 197

4.3 Results and Discussion 197

5 Future Work and Conclusion 199

References 200

Chapter 10 Arabic and English Typeface Personas Shima Nikfal Ching Y. Suen 203

1 Introduction 203

2 Literature Review of Typeface Personality Studies 204

3 Arabic Typeface Personality Traits 207

3.1 Research methodology 207

3.2 Statistical analyses of survey results 212

4 English Typeface Personality Traits 217

4.1 Research methodology 217

4.2 Statistical analyses of survey results 221

5 Summary of English Typefaces 225

6 Summary of Arabic Typefaces 226

7 Comparison of Both Studies 226

8 Conclusions and Future Work 227

References 228

Chapter 11 End-to-End Lexicon Free Arabic Speech Recognition Using Recurrent Neural Networks Abdelrahman Ahmedy Yasser Hifny Khaled Shaalan Sergio Toral 231

1 Introduction 231

2 Related Work 232

3 Arabic Speech Recognition System 233

3.1 Acoustic model 234

3.2 Language model 237

3.3 Decoding 237

4 Front-End Preparation 239

4.1 Converting the Arabic text to Latin (transliteration process) 239

4.2 Converting the transcription to alias 240

4.3 Speech features extraction 240

5 Experiments 241

5.1 The 8-hour experiment 241

5.2 The 8-hour results 242

5.3 The 1200-hour experiment 244

5.4 The 1200-hour results 245

6 Conclusion 245

References 246

Chapter 12 Bio-Inspired Optimization Algorithms for Improving Artificial Neural Networks: A Case Study on Handwritten Letter Recognition Ahmed A. Ewees Ahmed T. Sahlol 249

1 Introduction 249

2 Neural Networks and Bio-inspired Optimization Algorithms 252

2.1 Neural Networks (NNs) 252

2.2 Particle Swarm Optimization (PSO) 252

2.3 Evolutionary Strategy (ES) 252

2.4 Probability Based Incremental Learning (PBIL) 253

2.5 Moth-Flame Optimization (MFO) 253

3 Swarms Working Mechanism 255

4 The Proposed Approach 257

5 Experiments and Results 258

5.1 Dataset description 258

5.2 Evaluation criteria 259

5.3 Results and discussions 259

6 Conclusion and Future Work 264

References 265

Index 267