A Method for the Affixed Word Transliteration to the Balinese Script on the Learning Web Application

This research proposed a method for the affixed word transliteration to the Balinese Script since there has not been studied yet and it is important since the affixed word needs to be transliterated, inevitably. This research is one of the efforts to preserve digitally the endangered Balinese local language knowledge in Indonesia through the multi-discipline collaboration between Computer Science and Language discipline. The proposed method was taken care of two related aspects, i.e.; (1) A Latin root word has its related Balinese Script root word by using default or special transliteration rule; and (2) A Latin root word with a special transliteration rule for its Balinese Script root word, also need a special transliteration rule for its affixed word. This study was conducted on the pioneering web-based transliteration learning application, BaliScript, that receives the Latin text input and outputs the Balinese Script by using the Noto Serif Balinese font with its dedicated Balinese Unicode. Through the experiment, the proposed method gave the expected transliteration results, added a certain perspective, and strengthened the transliteration knowledge. Future work is to enhance and reuse this method on the mobile computing device, as a part of the Balinese Language ubiquitous learning that supports Balinese Language education, which is a mandatory local subject from the elementary school to the high school in Bali Province.


Introduction
The Balinese Script as an endangered local language (G. Indrawan, Paramarta, Agustini, & Sariyasa, 2018) in Indonesia affects its Latin-to-Balinese Script transliteration knowledge, as part of the transliteration research in general (Karimi, Scholer, & Turpin, 2011;Kaur & Singh, 2014). The preservation effort by the Bali Government has already been conducted through the Bali Governor Circular Letter (Bali Government, 2018) that follows up the Bali Governor Regulation (Bali Government, 2019). Also, from the previous preservation regulation, the Balinese Language (including all aspects of the Balinese Script) was considered and running as a mandatory local subject from elementary school to senior high school in Bali Province.
As the preservation effort of the language can be done by that governmental/political approach, multiple approaches should strengthen it and should have a greater impact. This research joined the effort through the technological approach, through the multi-discipline collaboration between Computer Science and Language discipline, by proposing a method for the affixed word transliteration to the Balinese Script since there has not been studied yet and it is important since the affixed word needs to be transliterated, inevitably. Moreover, this study was conducted on the pioneering web-based transliteration learning application, BaliScript, as part of the planned Balinese Language ubiquitous learning that supports Balinese Language education. The proposed method can be reused in the future for the mobile application as part of ubiquitous learning to give the transliteration knowledge for the next generations to come. As known, ubiquitous learning improves learning by providing learners with content and interaction anytime and anywhere (Hwang, Tsai, & Yang, 2008) by mobile and embedded computers and wireless networks in everyday life (Ogata, Matsuka, Bishouty, & Yano, 2009). This research proposes an unexposed method for the affixed word transliteration to the Balinese Script based on the supporting computer font Noto Serif Balinese (NSB). Also, this work advances the previous authors" work on the transliteration process by: (1) accommodating the rule from the Balinese Language, Script, and Literature Advisory Agency (Anom et al., 2009). This Bali Province government agency (Bali Government, 1992) carries out guidance and formulates programs for the maintenance, study, development, and preservation of the Balinese Language, Script, and Literature; (2) accommodating seventeen types of special words (G. Indrawan, Paramarta, & Agustini, 2019; G. Indrawan, Paramarta, et al., 2018) by using a particular table structure in the database which is avoiding hard-coded special word repositories in the learning application; (3) utilizing the Noto Serif Balinese font (Google, 2020b; The Unicode Consortium, 2020b, 2020a) with more developed and less bug than the Noto Sans Balinese font (Google, 2020a). Also, it uses dedicated Balinese

Research Article
Vol. 12 No.6 (2021), [2849][2850][2851][2852][2853][2854][2855][2856][2857] Unicode as a standard Balinese Script font on the computer system including mobile devices (so the method can be reused on the mobile application); and (4) maximizing the learning application, that used this method, for the delivery of the knowledge in the area of transliteration and also translation, at the same time, specifically for the Balinese Script, Indonesian, and English (see the next Figure 2). Overall, all of those advances were considered as the contribution of this work. This paper is organized into several sections, i.e.: Section 1 (Introduction) states the problem background related to the Balinese Script in general, and the Latin-to-Balinese Script transliteration in specific; Section 2 (Related Work) describes the related works in the area of the Latin-to-Balinese Script transliteration; Section 3 (Research Method) exposes the supporting transliteration computer font (which is the NSB font) with its dedicated Balinese Unicode, and the supporting rule-based algorithm for the affixed word transliteration to the Balinese Script; Section 4 (Result and Discussion) covers the analysis of the testing result; and finally, Section 5 (Conclusion and Future Work) consists of some important conclusion and future work points.

Related Work
As the Balinese Script adheres with its scriptio-continua writing style like its relatively close (geographically and culturally) Javanese Script (Widiarti & Pulungan, 2020), several related works on Latin-to-Balinese Script transliteration were conducted on the previous author"s works (Crisnapati et (Sudewa, 2003). It was also used on the developed robotic system that writing Balinese Script from Latin text input (Crisnapati et

Research Method
The proposed method for the affixed word transliteration to the Balinese Script was relied on two related aspects, i.e.; (1) A Latin root word has its related Balinese Script root word by using default or special transliteration rule; and (2) A Latin root word with a special transliteration rule for its Balinese Script root word, also need a special transliteration rule for its affixed word.
This section describes: (1) the supporting transliteration computer font (which is the NSB font) with its dedicated Balinese Unicode, as a standard Balinese Script font on the computer system including mobile devices; and (2) the supporting rule-based algorithm for the affixed word transliteration to the Balinese Script.
That example of the affixed word transliteration and several other cases (using prefix and/or suffix) should be handled by the proposed method. Algorithm 1 shows the pseudocode for the affixed word transliteration to the Balinese Script. Figure 1 shows the generated SQL instruction for the example of word_string = "aksarane" with word_ string_length = 8.
Based on the example, line 01 gave part of SQL instruction, i.e. "SELECT "aksarane" varian, 8 varianLength". Line 02 sets word_min = 3 which means searching in the database for the root words with a minimum of three characters. Line 03 sets the initial value of word_substring = "aksarane" for the first for iteration at line 04. Line 05 and 06 set the initial value of word_substring2 = "aksaran" with word_ substring2_length = 7 for the second for iteration at line 07. Line 08 gave part of SQL instruction, i.e. "UNION ALL SELECT "aksaran", 7". Line 09 updates the value of word_substring2 = "aksara" with word_ substring2_length = 6 for the second for iteration at line 07. The condition of word_substring2 ≥ word_min would break the second for iteration at line 07 and continue with the first for iteration at line 04. Figure 1 shows the complete variants of word_substring2 that would be JOIN with records in the database if column "kata" equal to those variants of word_substring2. Line 13 result was shown at the bottom section of Figure 1. Only the top record was used as the result since it has the biggest match level compare to the other records. For the sake of

Research Article
Vol.12 No.6 (2021), 2849-2857 clarity, the other records result was shown by commented line "LIMIT 1" of the constructed SQL instruction. Algorithm 1. The affixed word transliteration to the Balinese Script 01 insert word_string and its word_ string_length to SQL SELECT instruction 02 set word_min = 3 03 set word_substring = word_string 04 for word_substring and word_substring ≥ word_min, do 05 set word_substring2 = word_substring 06 update word_substring2 by removing its most left character 07 for word_substring2 and word_substring2 ≥ word_min, do 08 insert word_ substring2 and its word_ substring2_length to SQL UNION ALL SELECT instruction 09 update word_substring2 by removing its most right character 10 end 11 update word_substring by removing its most left character 12 end 13 retrieve kata and its related konvert from the database by using constructed SQL instruction with "records.is_exclude = 0" and "LIMIT 1" 14 update word_string by replacing its same string part (refer to kata) with its related konvert of kata 15 go to the next transliteration process to get the Balinese Script output from the word_string Figure 1. The example result from the fully constructed SQL instruction that was generated by the algorithm Finally, line 15 gave word_string = "\akßarane" from column value "konvert" for the next transliteration process to give the right result "ᬅᬓ᭄ᬱᬭᬦᬾ". As a note, the SQL instruction search in the database for all related words which have flag status is_excluded = 0. Also "konvert" was referred to as Balinese code in the next

The Testing
The testing of the proposed method was conducted on the web application that was run on Intel(R) Core(TM) i7-4600U CPU @ 2.09GHz platform with 8 GB RAM and Windows 8 64-bit Operating System. That web application was constructed by the web server of Apache 2.4.46 (Win64), the database server MySQL 10.4.18-MariaDB, and programming language PHP 7.4.16 combined with JavaScript. Also as one of the contributions of this work, this web learning application, that used this method, maximizing the learning experience through the delivery of the knowledge in the area of transliteration and also translation at the same time, specifically for the Balinese Script, Indonesian, and English ( Figure 2). The translation information comes from the columns from the database. For example, Indonesian translation for certain transliterated word can be seen in the column "indo" (Figure 1) Table 2 shows the testing transliteration cases on several affixed words that exist from the testing data set of the previous authors" work (G. . Those cases were also accompanied by their counterpart testing cases that comply with the rule from the Balinese Language, Script, and Literature Advisory Agency (Anom et al., 2009). All of those testing cases, along with the root words to be transliterated, consists of various affixed words (the bold words) that using a prefix and/or suffix. Akeh akśaranē, 47, luir ipun: akśara suara, 14, akśara wianjana, 33, akśara suara punika talěr dados pangangge suara, tur madrěwe suara kakalih, kawāśţanin: suara hrěswa miwah dīrgha.
Line Break section in (Sudewa, 2003). Its English: Many of those letters, 47, as follows vowels, 14, consonants, 33, those vowels also become vowel signs, and have two sounds, that was called: sound hrěswa and dīrgha. 1 The original testing cases from the previous authors" work testing data set (G. Indrawan Figure 3 shows the testing result of the proposed method for the affixed word transliteration on the transliteration learning web application BaliScript. The testing on the BaliScript was based on testing cases in Table 2. Figure 3(a) shows the input with various affixed words through the select input box 2 , while Figure 3(b) shows the transliteration output with a line break for each testing case (Table 2). Several aspects were analyzed related to the transliteration process that was successfully displayed by the Noto Serif Balinese (NSB) font with its dedicated Balinese Unicode.

Result and Analysis
Balinese Script writing uses the non-scriptio-continua style (Widiarti & Pulungan, 2020) which means the output should fully occupy all of the writing space without paragraph and line break. For ease of learning and also for the ease of visual analysis, this aspect could be accommodated in the future by a feature to switch between the scriptio-continua style (the traditional look) and the non-scriptio-continua style (the modern look). The non-scriptio continua style using Cascading Style Sheets (CSS) property white-space 3 with value pre-line.  The first affixed word, i.e. an Indonesian word "Pembangunan" (development) was constructed by a prefix "pem", a root word "bangun" (develop), and a suffix "an". The writing of "e" and its sound [ə] (Esling, 1999) complies with the rule from the Balinese Language, Script, and Literature Advisory Agency (Anom et al., 2009).

Research Article
Vol. 12 No.6 (2021), 2849-2857 The affixed word "Pembangunan" was successfully transliterated by the proposed method to the Balinese Script "ᬧᭂ ᬫ᭄ᬩᬗᬸ ᬦᬦ᭄" by using its Balinese code (see column "konvert" at the previous algorithm section) related to the existing root word "bangun" ("ᬩᬗᬸ ᬦᬦ᭄") in the database. As a note, consonant "n" ("ᬦ᭄") of word "Pembangunan" was transformed to "nda" ("ᬦ᭄ᬤ") since it was followed by a syllable "da" from word "Daerah". The appended form of "da" was positioned under the regular form of "na" ("ᬦ").
The second affixed word, i.e. a Balinese word "akśarane" (that alphabets) was constructed by a root word "akśara" (alphabets) and a suffix "ne". The root word "akśara" (Sudewa, 2003), as a variant word of "aksara" (Anom et al., 2009) with the same transliteration, is accommodated by the proposed method. Since its intended sound is [e] (Esling, 1999), to comply with the rule from the Balinese Language, Script, and Literature Advisory Agency (Anom et al., 2009), the writing of suffix "ne" should be changed to "nē". Based on that change, the affixed word "akśaranē" was successfully transliterated by the proposed method to the Balinese Script "ᬅᬓ᭄ᬱᬭᬦᬾ" by using its Balinese code related to the existing root word "akśara" ("ᬅᬓ᭄ᬱᬭ") in the database.
The third affixed word, i.e. a Balinese word "punika" (that) was constructed by a prefix "pu", and a root word "nika" (that). The affixed word "punika" was successfully transliterated by the proposed method to the Balinese Script "ᬧᬸ ᬦᬶ ᬓ" by using its Balinese code related to the existing root word "nika" ("ᬦᬶ ᬓ") in the database.
The fourth affixed word, i.e. a Balinese word "pangangge" (sign) was constructed by a prefix "pang", and a root word "angge" (use). The root word "angge" (Sudewa, 2003), as a variant word of "anggē" (Anom et al., 2009) with the same transliteration, is accommodated by the BaliScript. Rather than using its Balinese code related to the existing root word "angge" ("ᬳᬗ᭄ᬕᬾ") in the database, the affixed word "pangangge" was successfully transliterated by the proposed method to the Balinese Script "ᬧᬗᬗ᭄ᬕᬾ" by using its Balinese code related to the existing root word "ngangge" ("ᬗᬗ᭄ᬕᬾ") in the database (see Algorithm 1). As a note, syllable "pa" ("ᬧ") of word "pangangge" was transformed to "spa" ("ᬲ᭄ᬧ") since it followed a consonant "s" ("ᬲ᭄") from word "dados". The appended form of "pa" was positioned after the regular form of "sa" ("ᬲ").
The fifth affixed word, i.e. a Balinese word "madrěwe" (have/has) was constructed by a prefix "ma" and a root word "drěwe" (belonging). The root word "drěwe" (Sudewa, 2003), as a variant word of "drewē" (Anom et al., 2009) with the same transliteration, is accommodated by the BaliScript. The affixed word "madrěwe" was successfully transliterated by the proposed method to the Balinese Script "ᬫᬤᬺ ᬯᬾ" by using its Balinese code related to the existing root word "drěwe" ("ᬤᬺ ᬯᬾ") in the database.
The sixth affixed word, i.e. a Balinese word "kakalih" (both) was constructed by a prefix "ka" and a root word "kalih" (two). Rather than using its Balinese code related to the existing root word "kalih" in the database, the affixed word "kakalih" was successfully transliterated by the proposed method to the Balinese Script "ᬓᬓᬵᬮᬶ ᬄ" by setting the flag status "is_excluded" = 1 of the existing root word "kalih" ("ᬓᬵᬮᬶ ᬄ") in the database (see Algorithm 1). As a note, rather than using the Balinese Script "ᬓᬓᬮᬶ ᬄ" (Sudewa, 2003) as the ground truth for the transliteration result of the affixed word "kakalih", the proposed method used "ᬓᬓᬵᬮᬶ ᬄ" The seventh affixed word, i.e. a Balinese word "kawāśţanin" (be named) was constructed by a prefix "ka", a root word "wāśţa" (name), and a suffix "nin". The root word "wāśţa" (Sudewa, 2003), as a variant word of "wasta" (Anom et al., 2009) with the same transliteration, is accommodated by the BaliScript. The affixed word "kawāśţanin" was successfully transliterated by the proposed method to the Balinese Script "ᬓᬯᬱ᭄ᬝᬦᬶ ᬦ᭄" by using its Balinese code related to the existing root word "wāśţa" ("ᬯᬱ᭄ᬝ") in the database. As a note, rather than using the Balinese Script "ᬓᬯᬵᬱ᭄ᬝᬦᬶ ᬦ᭄" (Sudewa, 2003) as the ground truth for the transliteration result of the affixed word "kakalih", the proposed method used "ᬓᬯᬱ᭄ᬝᬦᬶ ᬦ᭄" (without "ᬵ" as the Balinese Vowel

Conclusion and Future Work
A method for the affixed word transliteration to the Balinese Script was proposed with the implementation on the web learning application by using Noto Serif Balinese font with its dedicated Balinese Unicode. All various cases were handled successfully by the proposed method. This research added a certain perspective and strengthened the transliteration knowledge, which is part of the Balinese culture preservation effort.
Future work to enrich and strengthen this transliteration knowledge is to enhance and reuse this method on mobile computing through the implementation on the mobile device, as a part of the planned Balinese Language ubiquitous learning that supports Balinese Language education.