Model for Translation of English Language Noun Phrases to Luganda

London Journal of Research in Computer Science and Technology
Volume | Issue | Compilation
Authored by Fagbolu Olutola Olaide , Wasike Azizi
Classification: NA
Keywords: NA
Language: English

In Uganda, Luganda is the most dominant local language spoken all over the country with over 6 million speakers thus, learning Luganda would enable English speaking foreign traders, tourists and Non- Governmental Organisations (NGOs) to get the best out of their dealings with the Luganda speakers. Leading Neural Machine Translation (NMT) Systems such as Google Translate (GT), Microsoft Translator and those based on Statistical Machine Translation (SMT) cannot satisfactorily support Noun Phrase translations between major and minor languages such as the English language and Luganda pair because of inadequate digital resources for minority languages. Machine Translation (MT) tools which are potentially affordable and   nvenient  anguage learning options cannot help majority language speakers to learn minority languages due to diversities in the syntax and semantic structures of the languages but cheaper and more effective MT approach between major and minor languages is Rule- based Machine Translation (RBMT) which involves harvesting a language pair’s linguistic knowledge. Design Science Research Methodology (DSRM) allows for continuous refinement of a translation model and its implementation through  rototyping techniques, Document Analysis and Focus Group Discussions to incorporate new translation rules into the model and use Holdout validation with Human Evaluation to test the model output. Deterministic Finite Automaton (DFA) automates the Noun Phrase Translation model, Java Formal Languages and Automata Package (JFLAP) designs the DFA while Python is the programming language. The bilingual dictionaries are implemented as Coma Separated Values (CSV) files, the Natural Language Tool Kit (NLTK) supports Natural Language Processing (NLP) tasks such as Part-Of-Speech (POS) tagging, parsing and tkinker develops a Graphical User Interface (GUI) for the MT application. Py2Exe creates an executable file from the python codes and Nullsoft Scriptable Install System (NSIS) builds the window installer for the application. The translation model does not cover complex noun phrases consisting of other phrases such as prepositional phrases and focuses on the commonest noun phrase pattern of the Pre-modifier + Head structure. Luganda as a noun centric language with 10 noun classes does not have standalone articles but exist as prefixes and noun modifiers are determined by noun’s grammatical number and its class. Evaluation results offers clear and unambiguous translation of English to Luganda noun phrase which invariably facilitate tutoring and teaching of Luganda. This research work aids the protection of both the language and the culture from a possible extinction. As a matter of fact, the dominancy of English in African societies has its advantages but it has damaged African local languages values and virtues that is language is culture therefore, if a language dies, culture dies and when a culture dies, its people die.

               

Model For Translation of English Language Noun Phrases to Luganda

Fagbolu Olutola Olaideα & Wasike Aziziσ

____________________________________________

ABSTRACT

In Uganda, Luganda is the most dominant local language spoken all over the country with over 6 million speakers thus, learning Luganda would enable English speaking foreign traders, tourists and Non-Governmental Organisations (NGOs) to get the best out of their dealings with the Luganda speakers. Leading Neural Machine Translation (NMT) Systems such as Google Translate (GT), Microsoft Translator and those based on Statistical Machine Translation (SMT) cannot satisfactorily support Noun Phrase translations between major and minor languages such as the English language and Luganda pair because of inadequate digital resources for minority languages. Machine Translation (MT) tools which are potentially affordable and convenient language learning options cannot help majority language speakers to learn minority languages due to diversities in the syntax and semantic structures of the languages but cheaper and more effective MT approach between major and minor languages is Rule-based Machine Translation (RBMT) which involves harvesting a language pair’s linguistic knowledge. Design Science Research (DSR) Paradigm (DSRM) allows for continuous refinement of  Translation Model and its implementation through prototyping techniques, Document Analysis and Focus Group Discussions to incorporate new translation rules into the model and use Holdout Validation with Human Evaluation to test the model output. Finite State Transducer  (FST) automates the Noun Phrase Translation Model, Java Formal Languages and Automata Package (JFLAP) designs the FST while Python is the programming language. The bilingual dictionaries are implemented as Comma Separated Values (CSV) files, the Natural Language Toolkit (NLTK) supports Natural Language Processing (NLP) tasks such as Part-Of-Speech (POS) tagging, parsing and tkinker develops a Graphical User Interface (GUI) for the MT application. cx_Freeze creates an executable file from the python codes and Nullsoft Scriptable Install System (NSIS) builds the window installer for the application. The Translation Model does not cover complex Noun Phrases consisting of other phrases such as Prepositional Phrases and focuses on the commonest Noun Phrase pattern of the Pre-modifier + Head structure. Luganda as a Noun centric language with 10 Noun classes does not have standalone Articles but exist as prefixes and Noun modifiers are determined by Noun’s grammatical number and its class. The MT system was evaluated for Fluency and Accuracy by a Luganda lecturer. Evaluation results showed that the majority of the translations contained all meaning represented in their corresponding English language Noun Phrases and were of flawless language fluency and thus offers clear and unambiguous translation of English to Luganda Noun Phrases which invariably facilitate tutoring and teaching of Luganda. This research work aids the protection of both the language and the culture from a possible extinction.

Keywords: Luganda, Translator, Noun phrases, Document Analysis, Focus Group Discussions, Minority Llanguage, Culture.

  1. INTRODUCTION

Luganda is a Bantu language widely spoken in Uganda and East African regions including Kagera in Tanzania, Bungoma in Kenya and Gatsata in Rwanda by artisans, farmers and traders and it was the primary official and instructional language in Ugandan primary schools until 1962 when English replaced it as Uganda’s official language. It is a tonal language with the high, low and falling tones [1][2].

In the 19th and 20th centuries, Luganda emerged as the language of dominance among the political class because Baganda chiefs were used by the British colonialists to pass across guiding regulations and information to Ugandans thus Luganda is the most predominantly spoken local language within and outside Uganda with well over six million speakers from Buganda and outside the kingdom. Central Uganda as the country’s heart of commerce, seat of administration and centre of excellence compelled the largest percentage of non-Luganda speakers to develop an interest in this language so as to effectively communicate and compete fairly well with people at the central [3].

With advances in technology, the world has now shrunk into a global village with Uganda as a microcosm. This globalization has brought about the establishment of international trade, multinational investments, bilateral relations and so on. In Uganda, there are numerous foreign companies and Non-Governmental Organizations (NGOs) offering their services to the populace but with their inability to effectively communicate in Luganda has made it difficult for NGOs, tourists, foreign and multinational companies to achieve their goals and objectives [4][5][6] [7].

Apart from all the aforementioned challenges, there is the danger of language extinction with approximately 7,000 endangered languages in this world is expected to die by the end of this century.  The diminishing number of native speakers and dominance of prevalent languages such as Mandarin, English, Chinese, Spanish, French, Portuguese, Arabic, Russian and Swahili have endangered almost all minority languages with Luganda inclusive, if language such as Latin is considered as dead language and Irish is moving towards extinction how much more Luganda. Nowadays, middle-class and elite parents ban speaking of their mother tongues at home and happily for them, teachers at school also strictly enforce the prohibition on these languages that is, preference to mother tongues no longer exist which might consequently lead to generations of Baganda born that are not fluent speakers of the language and if this trend should continue, Luganda will move to extinction. When a language dies, cultures die and consequently its people die and like Luganda, many other local languages are at risk of extinction but the digitization of these languages by using carefully designed linguistic computer translation models provides a ray of hope [8].

According to [2], development depends on knowledge and knowledge is delivered through language and for knowledge to be properly shared, language must be shared. To share language, the language can either be learned or translated and both can be achieved using Natural Language Processing (NLP) and Machine Translation (MT). Statistical Machine Translation (SMT) and Neural Machine Translation (NMT) tools such as Google Translate (GT), Microsoft Translator are successful stories of MT, however, they only support major languages such as English, French and Spanish leaving out 99.8% of the local languages like Luganda. To digitize the less resourced languages and solve limitations of SMT models, the more affordable Rule-Based Machine Translation approach to MT is the only option [9][10]. Consequently, the most advanced MT products cannot help major language speakers to learn less resourced languages such as Luganda. The solution is to develop RBMT models powered by Automata Theory that will produce translations that are good enough to facilitate the learning of these languages and indirectly ameliorate language barrier which causes undesirable effects such as NGOs failure to achieve their set goals, less enjoyable tourism experiences and hindrance to trade and commerce [4] [6] [7].

  1.  RESEARCH OBJECTIVES

2.1  General Objective

To develop a model for translation of English language Noun Phrases to Luganda.

2.2   Specific Objectives

  1. To analyze English language Noun Phrase syntactic rules and their Luganda counterparts.
  2. To design a model for translation of English language Noun Phrases to Luganda.
  3. To implement the model in (2).
  4. To Evaluate the model implementation in (3).
  1. SIGNIFICANCE OF RESEARCH
  1. Leading MT systems such as GT struggle to  satisfactorily translate English to Luganda because of the peculiarities in Luganda and often times the syntactical rules are distorted during translation hence need for different translation methodologies to  offer better translation results [10][11].
  2. Minority languages such as Luganda need digitization so as to enjoy international acceptability, enhance usability on the digital media and devices, and protect Luganda from a possible extinction [12][2].
  3. The study contributes to solving the problem of language barrier which has been detrimental to globalization through international trade, tourism and cross-border NGO service delivery [13][4][6][7].

  1. SCOPE OF THE STUDY.

The study is limited to development of a model for English Language Noun Phrases translation to Luganda and software application to validate the model

V.    RESEARCH METHODOLOGY

Design Science Research Process model  of the Design Science Research Paradigm guided the study. Document Analysis and Focus Group Discussions were used to analyse English language and Luganda Noun Phrase syntactic rules respectively. Purposive sampling was used to select 10 students and a lecturer from the College of Education, Open, Distance and E-Learning (CEODL) of Kampala International University, main campus. that and the MT system was evaluated by a Luganda lecturer using the Accuracy and Fluency parameters of human evaluation.The students and the Luganda lecturer developed the Luganda syntactic rules and evaluated the MT system respectively. Mathematical model for the translation of English Noun Phrases to Luganda is done using any of the three concepts in automata theory: Finite State Transducer (FST), Context Free Grammar (CFG) and Regular Expression (RegEx). FST and RegEx  were employed in this research work. Python  is the programming language of choice, English-Luganda bilingual dictionaries were implemented as Comma Separated Values (CSV) files and  NLTK supports tagging and parsing while Java Formal Languages and Automata Package (JFLAP) designs the FST. FSTis a finite-state machine that validates  a finite set of English Noun Phrases and outputs the appropriate Luganda  Noun Phrase translations. Generally, an FST is formally defined as a sextuple ⟨K,Σ,Γ,F,s,Δ⟩ where;

  1. K is a finite non-empty set of states.
  2. Σ is the input alphabet (a finite non-empty set of symbols).
  3. Γ is the output alphabet (a finite non-empty set of symbols).
  4. F is a set of Final States.
  5. S is an Initial State, an element of K.
  6. Δ is the state transition function.

For the FST in figure 1, the following are the specifications;

K={S0,S1,S2,S3,S4,S5,S6,S7,S8,S9,S10,S11,S12,   S13,S14,S15,S16,S17,S18,S19,S20,S21,S22}

Σ={NOUN,POSS,ADJ,ORDNUM,CARDNUM,QUANT,DEM,V-ING,V-ED,PREDET}

Γ={LNOUN,LPOSS,LADJ,LORDNUM,LCARDNUM,LQUANT,LDEM,LV-ING,LV-ED,LPREDET}

F={S1,S9,S12,S16,S17,S21}

s=S0

Δ=Relation from S×(Σ∪ϵ)  to S×(Γ∪ϵ)={

⟨⟨S0,NOUN⟩,⟨S0,LNOUN⟩⟩,⟨⟨S0,POSS⟩,⟨S0,LPOSS⟩⟩,⟨⟨S0,ADJ⟩,⟨S0,LADJ⟩⟩,

⟨⟨S0,PREDET⟩,⟨S0,LPREDET⟩⟩,⟨⟨S0,ORDNUM⟩,⟨S0,LORDNUM⟩⟩,

⟨⟨S0,CARDNUM⟩,⟨S0,LCARDNUM⟩⟩,⟨⟨S0,QUANT⟩,⟨S0,LQUANT⟩⟩,

⟨⟨S0,DEM⟩,⟨S0,LDEM⟩⟩,⟨⟨S0,V-ING⟩,⟨S0,LV-ING⟩⟩,⟨⟨S0,V-ED⟩,⟨S0,LV-ED⟩⟩,

⟨⟨S2,NOUN⟩,⟨S2,LNOUN⟩⟩,⟨⟨S2,ADJ⟩,⟨S2,LADJ⟩⟩,⟨⟨S2,ORDNUM⟩,⟨S2,LORDNUM⟩⟩,

⟨⟨S3,NOUN⟩,⟨S3,LNOUNJ⟩⟩,⟨⟨S4,NOUN⟩,⟨S4,LNOUN⟩⟩,⟨⟨S5,NOUN⟩,⟨S5,LNOUN⟩⟩,

⟨⟨S6,NOUN⟩,⟨S6,LNOUN⟩⟩,⟨⟨S6,ADJ⟩,⟨S6,LADJ⟩⟩,⟨⟨S7,NOUN⟩,⟨S7,LNOUN⟩⟩,

⟨⟨S7,ADJ⟩,⟨S7,LADJ⟩⟩,⟨⟨S8,NOUN⟩,⟨S8,LNOUN⟩⟩,⟨⟨S10,NOUN⟩,⟨S10,LNOUN⟩⟩,

⟨⟨S10,ADJ⟩,⟨S10,LADJ⟩⟩,⟨⟨S10,CARDNUM⟩,⟨S10,LCARDNUM⟩⟩,⟨⟨S11,NOUN⟩,⟨S11,LNOUN⟩⟩,

⟨⟨S13,NOUN⟩,⟨S13,LNOUN⟩⟩,⟨⟨S14,ADJ⟩,⟨S14,LADJ⟩⟩,⟨⟨S15,NOUN⟩,⟨S15,LNOUN⟩⟩,

⟨⟨S18,NOUN⟩,⟨S18,LNOUN⟩⟩,⟨⟨S18,DART⟩,⟨S18,ϵ⟩⟩,⟨⟨S19,NOUN⟩,⟨S19,LNOUN⟩⟩,

⟨⟨S19,ADJ⟩,⟨S19,LADJ⟩⟩,⟨⟨S20,NOUN⟩,⟨S20,LNOUN⟩⟩,⟨⟨S22,NOUN⟩,⟨S22,LNOUN⟩⟩}

The FST as seen in figure 1, is reactive to many peculiarities as seen in Luganda with 10 different Noun classes and it accepts an English language Noun Phrase and outputs the corresponding Luganda Noun Phrase.

The dotted arrows and their labels represent the Luganda Noun Phrase translations that occur at each final state using the Luganda syntax rules set {LR1, LR2, LR3, LR4, LR5, LR6, LR7, LR8, LR9, LR10, LR11, LR12, LR13, LR14, LR15, LR16, LR17, LR18, LR19, LR20, LR21, LR22}. The Luganda Noun Phrase syntax rules are re-writes from the English language Noun Phrase syntax rules set {R1, R2, R3, R4, R5, R6, R7, R8, R9, R10, R11, R12, R13, R14, R15, R16, R17, R18, R19, R20, R21, R22}.

The Flow Chart in Figure 2 shows the sequential steps in the process of Noun Phrase translation. The translation process starts with a user running the Noun Phrase Machine Translation System and inputting an English language Noun Phrase. The machine uses English language Noun Phrase linguistic rules to decide if the supplied phrase is a valid English language Noun Phrase. If invalid, the user has the opportunity of providing a grammatically correct Noun Phrase and if valid, a Luganda Noun Phrase is generated and displayed to the user. The output signifies the end of the translation process.

 Figure 3 is the Data Flow Diagram that shows the finer Machine Translation system processes. A user enters an English language Noun Phrase and the system tokenizes it and assigns parts-of-speech tags to each token. The system checks to see if it is a valid English language Noun Phrase and if invalid, the user has the option of rewriting the Noun Phrase. If the submitted phrase is a valid English language Noun Phrase, any Articles in the phrase are removed and the remaining tokens are translated word for word using bilingual dictionaries. The resulting Luganda tokens are manipulated to make credible Luganda translations of English language Noun Phrases which is eventually presented for viewing by the user.

The Use Case Diagram represents a user’s interaction with the Machine Translation system and shows the relationship between the user and the various Use Cases in which the user is involved as indicated in Figure 4.

The Sequence Diagram visually models the flow of logic within the system and facilitates both documentation and validation of logic as illustrated in figure 5. The various logic components are represented as arrow labels in the Sequence Diagram. The user provides the system with an English language phrase which is then processed by Machine Translation system. If the phrase is a valid English language Noun Phrase, a message confirming Noun Phrase validity is displayed to the user. Otherwise, an invalid Noun Phrase message is displayed for the user in which case the user will need to rewrite the phrase and submit it to the system for processing once more. If found to be a valid English language Noun Phrase upon processing, the phrase is translated into Luganda and translation subsequently displayed for the user.

Figure 1: Graphical representation of model for Translation of English Language Noun Phrases to Luganda

Figure 2: Flowchart of Model for Translation of English Noun Phrase(s) to Luganda

Figure 3: Data Flow Diagram of Model for Translation of English Noun Phrase(s) to Luganda

Figure 4: Use Case Diagram of Model for Translation of English Noun Phrase(s) to Luganda

Figure 5: Sequence Diagram of Model for Translation of English Noun Phrase(s) to Luganda

Design Science Research (DSR) Paradigm was used for this study which is known as learning through building, it facilitates the invention of technological artefacts with both universal and more contextualized local use-value to solve real-world problems by placing emphasis on the design, development and evaluation of applicable artefacts such as applications and methods.

The Design Science Research Process (DSRP)  model includes the following phases; Problem Identification and Motivation, Objectives of a solution, Design and Development, Demonstration, Evaluation, and Communication [18]. Figure 6 shows the various phases of DSRP. 

Figure 6: Design Science Research Process (DSRP) Model

5.1 Problem Identification and Motivation

This originates from the observation that existing Machine Translation systems do not provide good translations from major to less resourced languages.

5.2. Objectives of the Solution

The study aims at creating an effective model for the translation of English language noun phrases to Luganda with its Machine Translation applications.

5.3. Design and Development

5.3.1 Analysis of English Language and Luganda Noun Phrases

Document Analysis was employed so as to understand English language Noun Phrase structures and syntax. Focus Group Discussions were held with Luganda linguists to understand Luganda grammatical lexis and structures which invariably aided in understanding of how the English noun phrases are translated into Luganda.

5.3.2 Role of Prototyping Technique in the DSRP

Prototyping technique enabled the translation model to be continuously refined until acceptable translation results had been achieved and fully fitted in the proposed DSRP model. The approach was chosen because it was excellent at validating requirements, revealed critical design, reduced errors, identified performance-enhancing design changes and design refinement through simulated use. It was excellent at communication of design information, exploration of new design concepts and facilitated active learning as new knowledge was continuously obtained from the noun phrases [19].

5.3.3 Machine Translation Approach

Transfer-based Machine Translation method of the Rule-Based Machine Translation approach was used in the design of the translation model since it uses linguistic knowledge that enables deep analysis at the syntactic level which is most appropriate for translations of minority languages.

5.3.4 Model Design Methods

UML diagrams of the OOP modelling methodology were used to model the behaviour of the Noun Phrase translation. System modelling diagrams included Finite State Transducer, Flowchart, Use Case Diagram, Sequence Diagram and Data Flow Diagram.

5.4        Designing the Noun Phrase Translation model

FLAP 7.1 was used to model a Finite State Transducer that automates English language Noun Phrases translation to Luganda, applications such as translation systems are more concisely modelled, which allow engineers to solve problems using higher level programming languages. The English language noun phrase grammar was validated using Regular Expression (RegEx), the Luganda equivalence was used to generate the Luganda translation. RegEx was deployed for its integration in many programming tools and language, simplicity for example, given that simple search strings can be used to match anything accurately and efficiently.

5.5        Translation Model Implementation

Python programming language was used because of its unrivalled community support and useful pre-built libraries. The NLTK library supported NLP subtask and tkinter was used to design the GUI for the system.

NLTK was the library of choice as it has a large community of users with lots of learning resources and variety of algorithms which you can choose from.

Tkinter was the preferred tool for designing the application’s GUI as it is bundled with Python, open source and got an active community of both old and new users to offer help whenever need arises.

5.6        Demonstration

The phase involved demonstrating to the research team members and Doctoral Committee of School of Computing and Information Technology that the implementation of the translation model actually works.

5.7        Evaluation

The MT system was evaluated using a combination of Holdout Validation and Human Evaluation by a Luganda lecturer and linguist. Holdout validation divided the test data into two halves, one for testing of the system during development and the other for use by Luganda linguists during Human evaluation. Holdout validation was preferred because of its speed, simplicity and flexibility. Human Evaluation was involved because it is the gold standard of Machine Translation Evaluation and since translations are for meant for human use, humans are the best judges of quality [27][28][29].  

To judge the quality of the translations, the Fluency and Accuracy parameters of translation quality were used by the evaluator.

5.8        Communication

The research findings were communicated in this  Journal article.

  1. ANALYSIS AND DISCUSSION

The alphabet letters X and Q in the English language are replaced with ŋ and combination of letters ‘NY’ in the Luganda alphabet (Walifu y’oluganda) respectively and the combination ng’ replaces the letter ŋ because of its non-availability on standard keyboards.

The direction of translation is from English language (Source Language) to Luganda (Target Language), English language noun phrase syntax was used to create the appropriate Luganda noun phrase syntax. The commonest pattern for constructing English language noun phrases is the Pre-modifier + Head structure with the noun being the common head. Table 1 shows English language noun phrase rules and their corresponding Luganda patterns. 

   

English language Noun Phrase Syntactic Rules

Corresponding Luganda Noun Phrase Syntactic Rules

R1

NP = Article + N

LR1

NP = N

NP = N + Adj

R2

NP = Article + Adj + N

LR2

R3

NP = Article + Adj + Adj + N

LR3

NP = N + Adj + Adj

NP = N + OrdNum

R4

NP = Article + OrdNum + N

LR4

R5

NP = Poss+N

LR5

NP = N + Poss

NP = N + Poss + Adj

R6

NP = Poss+Adj+N

LR6

R7

NP = Poss+OrdNum+N

LR7

NP = N + Poss + OrdNum

NP = N + Quant

R8

NP = Quant+N

LR8

R9

NP = Quant+Adj+N

LR9

NP = N + Adj + Quant

R10

NP = Quant+ Adj + Adj + N

LR10

NP = N + Adj + Adj + Quant

R11

NP = CardNum+N

LR11

NP = N + CardNum

R12

NP = Dart + CardNum + N

 

NP = N + Poss

R13

NP = Dart + CardNum + Adj + N

LR13

NP = N + Adj + CardNum

R14

NP = preDet+N

LR14

NP = N + preDet

R15

NP = preDet+ Dart + N

NP = N + Adj + Adj + Quant

R16

NP = preDet+ Dart + Adj + N

LR16

NP = N + preDet + Adj

R17

NP = Dem + N

LR17

NP = N + Dem

R18

NP = Dem + CardNum + N

LR18

NP = N + Dem + CardNum

R19

NP = Dem + Adj + N

LR19

NP = N + Dem + Adj

R20

NP = Article + V-ed + N

LR20

NP = N + V-ed

R21

NP = Article + V-ing + N

LR21

NP = N + V-ing

R22

NP = N(pre-modifier) + N

LR22

NP = N+ N(pre-modifier)

Table 1: Common English language Noun Phrase syntactic rules and their corresponding Luganda Syntax

Luganda is Noun centric with its 10 noun classes creating noun phrase structures that are very different from their English language counterparts. For example, English Language Articles are independent words while their Luganda equivalents exist as prefixes attached to a Noun or Modifier. Furthermore, a Noun modifier or the Adjective such as  good has different forms in Luganda translation which are determined by the Noun’s class that they belong to and whether the Noun is singular or plural. Luganda is so peculiar and complex that a single Luganda word can contain so much syntactic information that it requires several English language words to translate it.

The model design deliberately excludes English language Articles as already seen, since they only exist in Luganda translations as prefixes and not standalone words. The translation quality achieved further substantiate that MT systems based on RBMT are more affordable and effective means of digitizing minority languages. It can be used to build language learning tools with which English language speakers can affordably and comfortably learn Luganda thereby enabling them to be more interactive to the Luganda speakers. The tool if used can also facilitate learning of Luganda and so help protect the language and culture of the Baganda from dilution and possible extinction.

  1. RECOMMENDATION AND CONCLUSION

In conclusion, the software artefact that translates English-Luganda Noun Phrases were developed, evaluated and deployed for use. Evaluation for Accuracy and Fluency revealed that the majority of the translations contained all meaning represented in their corresponding English language Noun Phrases and were of flawless language fluency . It is recommended that a study aimed at developing a model for translating entire English language sentences to Luganda would be great news for the digitization of Luganda. Future researchers can explore the efficiency and effectiveness of the model by examining alternatives to the NLTK library and data structure for bilingual dictionaries.

REFERENCES

  1. J. Tembe and B. Norton, Dreams and Realities: Developing Countries and the English Language, London: British Council, 2017.
  2. O. Fagbolu and B. Obalalu, “Smart Luganda language Translator,” in 7th International Conference on Natural Language Processing (NLP 2018), Dubai, 2018.
  3. T. Kalyegira, “Why i'm fed up with Luganda,” Daily Monitor, 14 June 2015. [Online]. Available:https://www.monitor.co.ug/Magazines/PeoplePower/Why-Im-fed-up-with-Luganda/689844-2750246-jw3489/index.html. [Accessed 12 April 2019].
  4. A. Ramnani, “The Role of Language in Shaping the International Tourism Experience of Student-travellers,” SJSU ScholarWorks, San Jose, 2012.
  5. D. Hendricks, “6 Silicon Valley Companies Going Global,” 8 Dec 2014. [Online]. Available:https://www.forbes.com/sites/drewhendricks/2014/12/08/6-silicon-valley-companies-going-global/#177c7ae56cf9. [Accessed 21 April 2019].
  6. E. Brynjolfsson, X. Hui and M. Liu, “Does Machine Translation Affect International Trade? Evidence from a Large Digital Platform,” 30 July 2018. [Online]. Available: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3210383.
  7. A. Crack, f. Hillary and w. Tesseur, “Many NGO workers on the ground don’t speak the local language – new research,” 8 August 2018. [Online]. Available: https://thecon versation.com/many-ngo-workers-on-the-ground-dont-speak-the-local-language-new-research-100845.
  8. A. Thompson, “How African Languages are Going Digital,” 15 may 2018. [Online]. Available:https://theculturetrip.com/africa/south-africa/articles/how-african-languages-are-going-digital/. [Accessed 12 April 2019].
  9. A. Hurskainen and J. Tiedemann, “Rule-based Machine Translation from English to Finnish,” in WMT, Copenhagen, 2017.
  10. F. Maulidiyah, “To Use or Not To Use Google Translate in English Language Learning,” Jurnal Linguistik Terapan, 2018.
  11. Google, “Language support,” 4 February 2019. [Online]. Available: https://cloud.google.com/ translate/docs/languages. [Accessed 20 April 2019].
  12. H. Davyth, “Breton- a digital language?,” 2017. [Online]. Available: www.dldp.eu/sites/ default/files/documents/DLDP_Breton-Report.pdf. [Accessed 26 April 2019].
  13. J. González and F. Casacuberta, “Phrase-based finite state models,” in Sixth Annual workshop, FSMNLP, Potsdam, 2007.
  14. P. Ken , T. Tuure and N. Björn , “Design science research genres: introduction to the special issue on exemplars and criteria for applicable design science research,” European Journal of Information Systems, pp. 129-139, 2018.
  15. B. D. Bala, “Design Science Research Methodology in Computer Science and Information System,” ResearchGate, 2019.
  16. V. Vaishnavi, W. Kuechler and S. Petter, “Design Science Research in Information Systems,” 2019. [Online]. Available: http://www.desrist.org/design-research-in-information-systems/. [Accessed 15 September 2019].
  17. V. M. Isabella, d. l. H. Retha , P. R. Shana and R. V. Karen, “Deploying Design Science Research in Graduate Computing Studies in South Africa,” in Twenty First Americas Conference on Information Systems, Puerto Rico, 2015.
  18. K. Peffers, T. Tuunanen, C. E. Gengler, M. Rossi, W. Hui, V. Virtanen and J. Bragge, “The design science research process: A model for producing and presenting information systems research,” 25 Feb 2006. [Online]. Available: https://www.researchgate.net/publication/228650671_The_design_science_research_process_A_model_for_producing_and_presenting_information_systems_research/citation/download. [Accessed 16 Sep 2019].
  19. C. Bradley , V. Vimal , L. Julie , A. David , J. Daniel , C. Richard , O. Kevin and W. Kristin , “Design prototyping methods: state of the art in strategies, techniques,and guidelines,” Design Science, 2017.
  20. I. Greene and M. Wangia, “Applying Finite-State Methods to the Swahili Language,” 2010. [Online]. Available: https://www.seas.upenn.edu/~cse400/CSE400_2009_2010/final_report/Greene_Wangia.pdf.
  21. G. Andrew and P. Drew , “Use Case Diagrams in Support:Deriving Understanding from the Picture,” Journal of Database Management, pp. 1-24, 2009.
  22. M. M. Nasir , Q. M. Asad , S. Sidra and G. Shumaila , “An Integration of UML Sequence Diagram with Formal Specification Methods— A Formal Solution Based on Z,” Journal of Software Engineering and Applications, pp. 372-383, 2015.
  23. H. A. Mohammed, “Journal of Management and Strategy,” Importance of Data Flow Diagrams and Entity Relationships Diagrams to Data Structures Systems Design in C++ “A Practical Example”, pp. 51-61, 2017.
  24. Robert, “NLTK vs. spaCy: Natural Language Processing in Python,” 27 April 2016. [Online]. Available: https://blog. Thedataincu bator.com/2016/04/nltk-vs-spacy-natural-language-processing-in-python/. [Accessed 6 August 2019].
  25. H. Fatuma, “The 6 Best Python GUI Frameworks for Developers,” 13 November 2017. [Online]. Available: https://blog. resellerclub.com/the-6-best-python-gui-frameworks-for-developers/. [Accessed 6 August 2019].
  26. Python.org, “24.1 Tkinter — Python interface to Tcl/Tk,” 6 July 2019. [Online]. Available: https://docs.python.org/2/library/tkinter.html. [Accessed 6 August 2019].
  27. C.-J. R. Marta and F. Mireia, “Study and Comparison of Rule-based and Statistical Catalan-Spanish Machine Translation Systems,” Computing and Informatics, vol. 31, p. 245–270, 2012.
  28. S. Mutuvi, “Introduction to Machine Learning Model Evaluation,” 16 April 2019. [Online]. Available:https://heartbeat.fritz.ai/introduction-to-machine-learning-model-evaluation-fa859e1b2d7f
  29. M. S. Mirjam and G. Donaj, Natural Language Processing - New Approaches and Recent Applications, IntechOpen, 2019



author

For Authors

Author Membership provide access to scientific innovation, next generation tools, access to conferences/seminars
/symposiums/webinars, networking opportunities, and privileged benefits.
Authors may submit research manuscript or paper without being an existing member of LJP. Once a non-member author submits a research paper he/she becomes a part of "Provisional Author Membership".

Know more

institutes

For Institutions

Society flourish when two institutions come together." Organizations, research institutes, and universities can join LJP Subscription membership or privileged "Fellow Membership" membership facilitating researchers to publish their work with us, become peer reviewers and join us on Advisory Board.

Know more

subsribe

For Subscribers

Subscribe to distinguished STM (scientific, technical, and medical) publisher. Subscription membership is available for individuals universities and institutions (print & online). Subscribers can access journals from our libraries, published in different formats like Printed Hardcopy, Interactive PDFs, EPUBs, eBooks, indexable documents and the author managed dynamic live web page articles, LaTeX, PDFs etc.

Know more