Design and Implementation of an Offline Android-based English to Igbo and Yoruba Machine Translation System

                         ABSTRACT

Machine Translation system is an automated system that translates text from a source language to target language. The source language is the main language upon which the target language is derived, while target language is the semantic equivalence of the source language. The source language and target language are described as the input and output to the MT system respectively. Machine translation is an area of applied research in human language processing in which ideas and techniques are drawn from different fields such as linguistics, computer science, artificial intelligence, translation theory and statistics. 

The extinction of indigenous languages (Yoruba, Igbo, and Hausa) is increasing unconsciously, thereby giving dominance of non-indigenous languages such as English. In a bid to preserve indigenous languages, Android-based English to Igbo and Yoruba Machine Translation System is developed to cater for this setback. Android platform was chosen over other platforms for the proposed system due its increase in popularity and its availability for wide range of people.

The English, Yoruba and Igbo text were collected from digital resource and printed copies through administrated questionnaires. The text was analysed and corrected using Takada desktop Tone marker application and sublime text editor. Comparisons between query that is, word to be translated, are  made  with  the  created  corpus using a set of stored rules.  The software design was carried out using the Unified Markup Language (UML) and implemented using Java programming language, Android  Software Development Kit such Andriod studio, Extensible Markup Language (XML) and for the database services, set of stored procedures was used to access the stored data and to query the database.

The system is an offline android-based English to Igbo and Yoruba machine translation system that translates text written in English to Yoruba and Igbo with its respective sound. Texts containing proper tone marked words that are related to school domain were considered as corpus for the system. Alpha testing, Beta testing, Real Time (Response Time) testing, Intelligibility Test, Accuracy Test and Word Error Analysis were used to evaluate the  system in order to determine system performance.

The android based platform for English to Igbo and Yoruba text translation aid the learning of Yoruba and Igbo language with ease by providing the right pronunciation of the words, by listening to the sound. It can be used in the educational sector to keep both languages alive by serving as learning tools to assist interested learners. After deploying the system, it was found to be accurate and intelligible with minimal errors. 

                  TABLE OF CONTENTS

Certification                   ii

Dedication                  iii

Acknowledgements          iv

Table of Contents                 v

List of Tables          x

List of Figures         xi

Abstract                  xiii

1INTRODUCTION

Background………………………………………………............1

Statement of  the problem…………...............….…..3

Justification…………………………………………………........3

Aim…………….............................…..........……….........4

Objectives of the study……....………..........……………4

Methodology………………………………………………....…...4

Scope  of study………………………………………………......5

     Contribution to Knowledge…........…...............5

Thesis layout……………………………………………........….5

2LITERATURE REVIEW

2.1    Introduction………............................……………7

2.2.  Brief history of Yoruba and Igbo.                            Language....................................................7 

2.3. Related work…………….........................….….….8

       2.3.1 The difference between the                                    Phonemes of Igbo and Yoruba……..….8

2.4Basic concepts of Mobile platform and its            operating system….……........................……10

2.5Basic concepts of Machine Translation               System…………………..........…….......................11

2.6Approaches in Machine translation.……...12

2.7Related Works…………………………..……......…...16

        2.7.1   Android-Based Machine                                          Translation System…….........….........16

        2.7.2    Web Based Machine Translation                           System……………….............................20

2.8    Summary……………….............…....................25

3.METHODOLOGY

3.1       Introduction………….................................26

3.2       Software Development Model…......……26

3.3       Planning...............................…………………27

3.3     Requirement Specification......………………28

         3.3.1   Functional Requirements......……..28

         3.3.2   Non Functional Requirements.…..28

                    3.3.2.1    User Requirement…......…29

                    3.3.2.2   System Requirement……..29

3.4       Analysis and Design…........………….........30

  3.4.1System Specification.................30

          3.4.2 The assumptions on which the                               system is built…..........…....….....…...30

          3.4.3 System Design………………………....……30

     3.4.3.1    User Interface......…..…….31

                     3.4.3.2    Use Case Diagram……....31

                     3.4.3.3    Sequence Diagram……....32

                     3.4.3.4    Flowchart……………………….33

                     3.4.3.5   Activity Diagram……...…...35

                     3.4.3.6  Class Diagram…..……...……36

3.5     Database Design…..…......................………37

3.6     Design Model………………………………......…...41

3.7     Mobile app framework…………………..........41

3.8     System Implementation and Testing…..41

4SYSTEM IMPLEMENTATION

4.1 Introduction…………………………………………...…43

4.2 System Implementation…...................….…43

         4.2.1 Graphics User Interface….......…......44

         4.2.2Text-Text Service………............…..44

         4.2.3Audio API Integration………..........44

         4.2.4Software tools, Framework and                              Toolkits utilized…………...............45

        4.2.5.Implementation Environment…..…...46

        4.2.6. Implementation Details…………...……47

4.3 Result and Discussion………………………....……49

4.4     Challenges and Solution proffered……….53

          4.4.1Challenges………………………………....53

          4.4.2Solution proffered……………….......53

4.5System Evaluation…………………………............53

4.5.1Alpha Testing…………........…….......54

        4.5.2. Beta Testing………………………………….…54

        4.5.3. Response Time………………................54

                 4.5.3.1Response Time Evaluation...57 

        4.5.4Intelligibility Test…………..............57

4.5.4.1     Experiment………………..……...58

4.5.4.2    Intelligibility Evaluation.……59

                 4.5.4.3    Results of Intelligibility                                             test.....................................59

                 4.5.4.4 Analysis of Intelligibility                                           test........................................60

         4.5.5 Accuracy Test………………...............….61

                   4.5.5.1  Results of Accuracy Test...61

        4.5.6Word Error Rate (WER).…........….62

5CONCLUSION AND RECOMMENDATION

5.1     Summary………………………………………......……65

5.2   Conclusion…………………………………………........65

5.3 Recommendation and Future work…………………………............……………...........…….65

REFERENCES……………..…………………………………..…67

Appendix A: Programming Codes………………....71

Appendix B: Literature Review Table………………88

Appendix C: Administered Questionnaire………92

                     

                        LIST OF TABLES

4.0: Illustrates the Service Respond Time of the proposed system………………...………...............……56

4.1: Shows the four point scale used for Intelligibility testing…………..……………….....…...……58

4.2: Shows the WER analysis of the test data………….….……………………………................…..…..63

               

                        LIST OF FIGURES

2.0 Direct MT Architecture………….….............…..15

2.1 Interlingua MT Architecture with 2 languages....…...................................................16

2.2: Web based major tools……………………..….....20

2.3: Translation process for English to Yoruba MT…………………………………………..….........................23

3.0: Iterative SDLC Model………………………........…27

3.1:  Use Case Diagram……………………………….……32

3.2: Sequence Diagram…………………………......……33

3.3: Flowchart Diagram…………………………......…..34

3.4: Activity Diagram……………………………….…..…..35

3.5: Class Diagram……………………………..........…..36 

3.6: English Database……………....................……38

3.7: Igbo Database……………………………………...……39

3.8: Yoruba Database……………………………….....….40

3.9: Interlingua Approach for Igbo….....…....…..41

3.10: Mobile App Framework Structure………...42

4.0: Welcome Page….………………………………..….…48

4.1: Screenshot of the word “Professor”……....50

4.2: Screenshot of the word “Lecturer”….....….51

4.3: Screenshot of the word “Lecturer”….......…52

4.4: Illustrates the offline performance evaluation of the system…................................57

4.5: shows the percentage of intelligibility testing scores given by evaluators…….............60

4.6: shows the percentage of accuracy testing scores given by evaluators……….…....................62

4.7: shows the distribution of word errors..…..64

A.1: The Java code snippet for the Implementation…………………………………………....…..76

A.2: The Java code snippet for the Database..  …………………………………......................................... 87

B.1: Literature Review Table…………………......……91

C.1: Administrated Questionnaires…………...……95