| { | |
| "paper_id": "O07-6001", | |
| "header": { | |
| "generated_with": "S2ORC 1.0.0", | |
| "date_generated": "2023-01-19T08:08:32.368346Z" | |
| }, | |
| "title": "Modeling Taiwanese Southern-Min Tone Sandhi Using Rule-Based Methods", | |
| "authors": [ | |
| { | |
| "first": "Un-Gian", | |
| "middle": [], | |
| "last": "Iunn", | |
| "suffix": "", | |
| "affiliation": { | |
| "laboratory": "", | |
| "institution": "National Taiwan University", | |
| "location": {} | |
| }, | |
| "email": "" | |
| }, | |
| { | |
| "first": "Kiat-Gak", | |
| "middle": [], | |
| "last": "Lau", | |
| "suffix": "", | |
| "affiliation": {}, | |
| "email": "" | |
| }, | |
| { | |
| "first": "Hong-Giau", | |
| "middle": [], | |
| "last": "Tan-Tenn", | |
| "suffix": "", | |
| "affiliation": {}, | |
| "email": "" | |
| }, | |
| { | |
| "first": "Sheng-An", | |
| "middle": [], | |
| "last": "Lee", | |
| "suffix": "", | |
| "affiliation": { | |
| "laboratory": "", | |
| "institution": "National Taiwan University", | |
| "location": {} | |
| }, | |
| "email": "" | |
| }, | |
| { | |
| "first": "Cheng-Yan", | |
| "middle": [], | |
| "last": "Kao", | |
| "suffix": "", | |
| "affiliation": { | |
| "laboratory": "", | |
| "institution": "National Taiwan University", | |
| "location": {} | |
| }, | |
| "email": "[email protected]" | |
| } | |
| ], | |
| "year": "", | |
| "venue": null, | |
| "identifiers": {}, | |
| "abstract": "A sizable corpus of Taiwanese text in Latin script has been accumulated over the past two hundred or so years. However, due to the special status of Taiwan, few people can read these materials at present. It is regrettable that the utilization of these plentiful materials is very low.", | |
| "pdf_parse": { | |
| "paper_id": "O07-6001", | |
| "_pdf_hash": "", | |
| "abstract": [ | |
| { | |
| "text": "A sizable corpus of Taiwanese text in Latin script has been accumulated over the past two hundred or so years. However, due to the special status of Taiwan, few people can read these materials at present. It is regrettable that the utilization of these plentiful materials is very low.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Abstract", | |
| "sec_num": null | |
| } | |
| ], | |
| "body_text": [ | |
| { | |
| "text": "This paper addresses problems raised in the Taiwanese Southern-Min tone sandhi system by describing a set of computational rules to approximate this system, as well as the results obtained from its implementation. Using the romanized Taiwanese Southern-Min text as source, we take the sentence as the unit, translate every word into Chinese via an online Taiwanese-Chinese dictionary (OTCD), and obtain the part-of-speech (POS) information from the Chinese Electronic Dictionary (CED) made by the Chinese Knowledge and Information Processing (CKIP) group of Academia Sinica. By using the POS data and tone sandhi rules based on linguistics, we then tag each syllable with its post-sandhi tone marker. Finally, we implement a Taiwanese Southern-Min tone sandhi processing system which takes a romanized sentence as an input and then outputs the tone markers.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "Our system achieves 97.39% and 88.98% accuracy rates with training and test data, respectively. Finally, we analyze the factors influencing error for the purpose of future improvement.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "Taiwanese is often used in daily life in Taiwan, but written Taiwanese is less common by far. Even so, the history of written Taiwanese stands at well over a century [Tiunn 2001 ]. At present, there are several dozen if not more than a hundred proposed phonetic and writing systems for Taiwanese [Iunn and Tiunn 1999] . The orthography adopted by this article is Peh-oe-ji (POJ, \u767d\u8a71\u5b57, also known as Latinized Taiwanese or Missionary Romanization System for Taiwanese).", | |
| "cite_spans": [ | |
| { | |
| "start": 166, | |
| "end": 177, | |
| "text": "[Tiunn 2001", | |
| "ref_id": "BIBREF9" | |
| }, | |
| { | |
| "start": 296, | |
| "end": 317, | |
| "text": "[Iunn and Tiunn 1999]", | |
| "ref_id": "BIBREF4" | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Background and Motivation", | |
| "sec_num": "1.1" | |
| }, | |
| { | |
| "text": "Under the auspices of the National Museum of Taiwanese Literature, the Department of Taiwanese Literature of Cheng Kung University carried out a project titled \"The Collection and Cataloging of Taiwanese Peh-oe-ji Literature Data\" (CCTPLD). Although many texts have already been lost due to the alternation of political status, this project nevertheless revealed nearly 2,000 POJ books and periodicals, with publication sites spread over Taiwan, Xiamen (Amoy), Shanghai, Guangzhou (Canton), Hong Kong, Singapore, the Philippines, London, Japan, and beyond. The amount of publishing peaked in the 1950's and 60's [Iunn and Tan-Tenn, unpublished] . The scope covers both formally published books and periodicals as well as non-published items such as personal letters and medical charts. Later on, the government, citing supposedly detrimental effects of POJ on Mandarin promotion, banned its use and thus caused the rapid decline of this practice.", | |
| "cite_spans": [ | |
| { | |
| "start": 612, | |
| "end": 644, | |
| "text": "[Iunn and Tan-Tenn, unpublished]", | |
| "ref_id": null | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Background and Motivation", | |
| "sec_num": "1.1" | |
| }, | |
| { | |
| "text": "We hope that the extant materials collected by the above-mentioned CCTPLD project can be accessed by more people, as well as contribute to both basic and applied Taiwanese research. As most people nowadays are not familiar with Latinized Taiwanese, use of state-of-the-art text-to-speech technology would enhance the value of these materials to the general public.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Background and Motivation", | |
| "sec_num": "1.1" | |
| }, | |
| { | |
| "text": "Tone sandhi represents a challenging problem to be solved before one can successfully transform the written Taiwanese text to its natural speech-like tonal contour. This is because the written form of Latinized Taiwanese represents the tones as \"basic tones\", the tones of syllables when they are pronounced in isolation. At the level of the word, all syllables except the last one are usually pronounced differently (that is, they manifest tone sandhi). At the level of a whole sentence, in most situations only the last syllables next to the boundary of the phrases or structural markers are read as basic tones, the others being read as sandhi tones. In fact, besides the \"regular tone sandhi\" mentioned above, there are still several other kinds of tone sandhi phenomena which will be discussed in detail later.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Background and Motivation", | |
| "sec_num": "1.1" | |
| }, | |
| { | |
| "text": "We will first formulate the sandhi rules, which are the key to correct pronunciation and the core issue of this paper. The input of our experiment mainly consists of the data collected by the CCTPLD project; these data are processed by our sandhi system to produce sandhi-marked final outputs. Due to the lack of tagged data, we adopt the rule-based model, not the statistical model in this experiment. Figure 1 describes the skeleton of our system, and the webpage http://iug.csie.dahan.edu.tw/nmtl/dadwt/ demonstrates the results. Three of the authors who are native Taiwanese speakers evaluated the outputs for their accuracy. low), 3 (low), 4 (middle short), 5 (low rising), 7 (middle flat), and 8 (high short). The tone pitch is described within the parentheses. Please refer to the following examples for tone diacritics. In this paper, all examples are written in Taiwanese. For the sake of apprehensibility, we also add the Mandarin and English translations.", | |
| "cite_spans": [], | |
| "ref_spans": [ | |
| { | |
| "start": 403, | |
| "end": 411, | |
| "text": "Figure 1", | |
| "ref_id": "FIGREF0" | |
| } | |
| ], | |
| "eq_spans": [], | |
| "section": "Background and Motivation", | |
| "sec_num": "1.1" | |
| }, | |
| { | |
| "text": "Tone sandhi is a very important characteristic of Taiwanese. At the word level, the last syllable is usually pronounced as basic tone and the others as sandhi tones. In example (1), the underlined syllables are pronounced as basic tones, the others as sandhi tones:", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Sound generator program", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "(1) t\u00e2i [\u53f0, \"platform\"] T\u00e2i-g\u00ed [\u53f0\u8a9e, \"Taiwanese language\"] T\u00e2i-g\u00ed-b\u00fbn [\u53f0\u8a9e\u6587, \"written Taiwanese\"] T\u00e2i-g\u00ed b\u00fbn-ha\u030dk [\u53f0\u8a9e\u6587\u5b78, \"Taiwanese literature\"]", | |
| "cite_spans": [ | |
| { | |
| "start": 8, | |
| "end": 23, | |
| "text": "[\u53f0, \"platform\"]", | |
| "ref_id": null | |
| }, | |
| { | |
| "start": 31, | |
| "end": 57, | |
| "text": "[\u53f0\u8a9e, \"Taiwanese language\"]", | |
| "ref_id": null | |
| }, | |
| { | |
| "start": 69, | |
| "end": 95, | |
| "text": "[\u53f0\u8a9e\u6587, \"written Taiwanese\"]", | |
| "ref_id": null | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Sound generator program", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "T\u00e2i-g\u00ed b\u00fbn-ha\u030dk-s\u00fa [\u53f0\u8a9e\u6587\u5b78\u53f2, \"history of Taiwanese literature\"] At the level of the syllable or the word, tone sandhi may manifest itself in at least the following several ways: (a) Normal sandhi: using reduplicated syllables as examples (the numbers within parentheses are reading tones).", | |
| "cite_spans": [ | |
| { | |
| "start": 19, | |
| "end": 61, | |
| "text": "[\u53f0\u8a9e\u6587\u5b78\u53f2, \"history of Taiwanese literature\"]", | |
| "ref_id": null | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Sound generator program", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "(2) (i) tone 1 \u2192 tone 7: \"chheng-chheng\" (7,1) [\u6e05\u6e05, \"clear\"] (ii) tone 7 \u2192 tone 3: \"ch\u0113ng-ch\u0113ng\" (3,7) [\u975c\u975c, \"quiet\"] (iii) tone 3 \u2192 tone 2: \"chhi\u00f2-chhi\u00f2\" (2,3) [\u7b11\u7b11, \"smiley\"] (iv) tone 2 \u2192 tone 1: \"l\u00e9ng-l\u00e9ng\" (1,2) [\uf92e\uf92e, \"cold\"] (v) tone 5 \u2192 tone 7 or 3 (northern Taiwan): \"\u00e2ng-\u00e2ng\" (7/3,5) [\u7d05\u7d05, \"red\"] (vi) tone 4 \u2192 tone 8 (-p/t/k) or 2 (-h): like \"sip-sip\" (8,4); [\u6fd5\u6fd5, \"moist\"] \"khoeh-khoeh\" (2,4) [\u64c1\u64e0, \"crowd\"] (vii) tone 8 \u2192 tone 4 (-p/t/k) or 3 (-h): like \"ti\u030dt-ti\u030dt\" (4,8) [\u76f4\u76f4, \"straight\"];", | |
| "cite_spans": [ | |
| { | |
| "start": 47, | |
| "end": 60, | |
| "text": "[\u6e05\u6e05, \"clear\"]", | |
| "ref_id": null | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Sound generator program", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "\"jo\u030dah-jo\u030dah\" (3,8) [\u71b1\u71b1, \"hot\"] (b) Following sandhi: this pattern generally occurs with pronouns or the suffix of names. The tone pitch depends on that of the preceding syllable and is either tone 1 (high), 3 (middle), or 7 (low).", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Sound generator program", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "(3) (i) \"A-eng--a\" (7,1,1) [\u963f\u82f1, a personal name] (the second \"a\" is a suffix) (ii) \"g\u00f3a l\u00e2i kh\u00f2a\u207f --i\" (1,7/3,3,3) [\u6211\uf92d\u770b\u4ed6, \"I come to see him/her\"] (the basic tone of \"i\"[\u4ed6, \"(s)he\"] is tone 1) (iii) \"h\ue029 --l\u00ed\" (7,7) [ \u7d66\u4f60, \"give you\" ] (the basic tone of \"l\u00ed\"[\u4f60, \"you\"] is tone 2) (c) Neutral sandhi: the syllable immediately preceding the neutral sandhi (marked orthographically with double hyphens same as (b)) is read as basic tone, and the tones of the neutral sandhi are pronounced softly as if they were tone 3 or tone 4.", | |
| "cite_spans": [ | |
| { | |
| "start": 137, | |
| "end": 146, | |
| "text": "him/her\"]", | |
| "ref_id": null | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Sound generator program", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "(4) (i) \"T\u00e2n--sian-si\u207f\" (5,3,3) [\u9673\u5148\u751f, \"Mr. T\u00e2n\"] (the original tones of \"sian-si\u207f \"[\u5148\u751f, \"Mr.\"] are tone 7 and tone 1) (ii) \"ki\u00e2\u207f--chhut-l\u00e2i\" (5,4,3) [\u8d70\u51fa\uf92d, \"walk out\"] (the original tones of \"chhut-l\u00e2i\" [\u51fa\uf92d, \"out\"] are tone 8 and tone 5) (d) Double sandhi: this pattern mostly appears in syllables ending in the glottal stop (-h) and having tone 4. The normal sandhi rules are applied twice in sequence (i.e. tone 4 \u2192 tone 2 \u2192 tone 1):", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Sound generator program", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "(5) (i) \"beh tha\u030dk-chu\" (1,4,1) [\u8981\uf95a\u66f8, \"want to read books\"] (\"beh\" [\u8981, \"want\"] is tone 4, but rather than becoming tone 2, it becomes tone 1) (ii) \"kh\u00ec g\u014da-kh\u00e1u\" (1,3,2) [\u53bb\u5916\u9762, \"go outside\"] (\"kh\u00ec\"[\u53bb, \"go\"] is tone 3, but rather than becoming tone 2, it becomes tone 1) (e) Pre-\u00e1 sandhi: the syllables before \u00e1 do not follow normal sandhi rules unless they are tone 1 or 2.", | |
| "cite_spans": [ | |
| { | |
| "start": 67, | |
| "end": 78, | |
| "text": "[\u8981, \"want\"]", | |
| "ref_id": null | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Sound generator program", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "(6) (i) tone 1 \u2192 tone 7: \"sun-\u00e1\" (7,2) [\u59ea\u5b50, \"nephew\"] (ii) tone 2 \u2192 tone 1: \"chh\u00e1u-\u00e1\" (1,2) [\u5c0f\u8349, \"grass\"] (iii) tone 3 \u2192 tone 1: \"t\u00e0\u207f-\u00e1\" (1,2) [\u6524\u4f4d, \"stall\"] (iv) tone 4 \u2192 tone 8 (-p/t/k) or tone 1 (-h): \"tek-\u00e1\" (8,2) [\u7af9\u5b50, \"bamboo\"] \"thih-\u00e1\" (1,2) [\u9435,\"iron\"] (v) tone 5 \u2192 tone 7: \"l\ue027-\u00e1 \" (7,2) [\uf932\u5b50, \"oven\"]", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Sound generator program", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "(vi) tone 7 does not change: \"ph\ue029-\u00e1\" (7,2) [\u7c3f\u5b50, \"tablet\"] (vii) tone 8 \u2192 tone 4 (-p/t/k) or tone 7 (-h): \"chha\u030dt-\u00e1\" (4,2) [\u8cca, \"thief\"] \"hio\u030dh-\u00e1 \" (7,2) [\uf96e, \"leaf\"] (f) Triplicate sandhi: the first syllable of triplicated words does not follow normal sandhi rules unless it is of tone 2, 3, or 4:", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Sound generator program", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "(7) (i) tone 1 \u2192 tone 5: like \"chheng-chheng-chheng\" (5,7,1) [\u6e05\u6e05\u6e05, \"very clear\"] (ii) tone 2 \u2192 tone 1: like \"\u00fan-\u00fan-\u00fan\" (1,1,2) [\u7a69\u7a69\u7a69, \"very stable\"] (iii) tone 3 \u2192 tone 2: like \"h\u00e8ng-h\u00e8ng-h\u00e8ng\" (2,2,3) [\u8208\u8208\u8208, \"very interesting\"]", | |
| "cite_spans": [ | |
| { | |
| "start": 61, | |
| "end": 80, | |
| "text": "[\u6e05\u6e05\u6e05, \"very clear\"]", | |
| "ref_id": null | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Sound generator program", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "(iv) tone 4 \u2192 tone 8 (-p/t/k) or tone 2 (-h): like \"sip-sip-sip\" (8, 8, 4) [\u6fd5\u6fd5\u6fd5, \"very humid\"] \"bah-bah-bah\" (2,2,4) [\u8089\u8089\u8089, \"very fat\"] (v) tone 5 \u2192 (similar to) tone 5: like \"k\u00f4a\u207f-k\u00f4a\u207f-k\u00f4a\u207f\" (5,7/3,5)", | |
| "cite_spans": [ | |
| { | |
| "start": 65, | |
| "end": 68, | |
| "text": "(8,", | |
| "ref_id": null | |
| }, | |
| { | |
| "start": 69, | |
| "end": 71, | |
| "text": "8,", | |
| "ref_id": null | |
| }, | |
| { | |
| "start": 72, | |
| "end": 74, | |
| "text": "4)", | |
| "ref_id": null | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Sound generator program", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "[\uf92e\uf92e\uf92e, \"very cold\"]", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Sound generator program", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "(vi) tone 7 \u2192 (similar to) tone 5: like \"ch\u0113ng-ch\u0113ng-ch\u0113ng\" (5,3,7)", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Sound generator program", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "[\u975c\u975c\u975c, \"very quiet\"]", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Sound generator program", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "(vii) tone 8 \u2192 (similar to) tone 5: like \"ti\u030dt-ti\u030dt-ti\u030dt\" (5,4,8) [\u76f4\u76f4\u76f4, \"very straight\"]", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Sound generator program", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "\"pe\u030dh-pe\u030dh-pe\u030dh\" (5,3,8) [\u767d\u767d\u767d, \"very white\"] (g) Rising sandhi: this pattern usually occurs on loanwords from Japanese; the sandhi tone is similar to tone 5.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Sound generator program", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "(8) \"\u014fai-siak-ch\u00f9\" (5,8,3) [\u767d\u896f\u886b, \"white shirt\" ] \"kh\u0103n-p\u00e1ng\" (5,2) [\u770b\u677f, \"signboard\"]", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Sound generator program", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "\"h\u0103n-t\ue023-l\u00f9 \"(5,1,3) [\u65b9\u5411\u76e4, \"steering wheel\"]", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Sound generator program", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "We collate the above sandhi phenomena in Table 1 . [Lin and Chen 1999] describes an early sandhi system. Users input Chinese texts, and the system outputs Taiwanese texts with pronunciation. The corpus is news reports in Chinese. They used the word segmentation and tagged data from the CKIP group and the Taiwanese-Chinese dictionary from Robert Cheng to map the Chinese news into Taiwanese (in both Han and Latin scripts). The sandhi rules applied were as follows: a) pronounce the last syllable at the end of a sentence as basic tone; b) pronounce the syllable before the particle \u00ea as basic tone; c) pronounce the last syllable of a noun as basic tone; d) pronounce other syllables as normal sandhi tones. An accuracy rate of 82.53% was reported. However, as the system did not take Taiwanese as input, word order and semantic ambiguities were not taken into account when converting, the translation was not quite native-like.", | |
| "cite_spans": [ | |
| { | |
| "start": 51, | |
| "end": 70, | |
| "text": "[Lin and Chen 1999]", | |
| "ref_id": "BIBREF7" | |
| } | |
| ], | |
| "ref_spans": [ | |
| { | |
| "start": 41, | |
| "end": 48, | |
| "text": "Table 1", | |
| "ref_id": "TABREF1" | |
| } | |
| ], | |
| "eq_spans": [], | |
| "section": "Sound generator program", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "[ Liang et al. 2004 ] is a recent text-to-speech system for Taiwanese Southern-Min. Its input was a large corpus of Chinese news texts, but sentences longer than 20 syllables were removed. It utilized a dictionary to convert the Chinese text into Taiwanese Southern-Min, followed by word segmentation, phonetic marking, and rule-based sandhi processing to generate speech files. Due to the size of the corpus, only the first 200 sentences generated were evaluated by two Taiwanese-speaking experts. The accuracy rates were 97% for word segmentation, 89% for pronunciation marking, and 65% for rule-based sandhi processing.", | |
| "cite_spans": [ | |
| { | |
| "start": 2, | |
| "end": 19, | |
| "text": "Liang et al. 2004", | |
| "ref_id": "BIBREF6" | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Historial Review", | |
| "sec_num": "1.3" | |
| }, | |
| { | |
| "text": "Compared with the above systems, our approach has some major differences: a balanced Taiwanese corpus for both literary and non-literary sources (about 50% each) was prepared; no translation from Chinese to Taiwanese; and no limits for length of sentences. In addition, because the text is written in Latinized script, we do not need to manipulate word segmentation and phonetic marking. However, compared to text with Han character script, there is a more rigorous challenge to deal with homonymy, especially with monosyllabic words.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Historial Review", | |
| "sec_num": "1.3" | |
| }, | |
| { | |
| "text": "The input data of our system are from the CCTPLD project. Following POJ orthography, syllables of a word are joined by hyphens, and the words are separated with spaces.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Data", | |
| "sec_num": "2.1" | |
| }, | |
| { | |
| "text": "We select parts of four sources as training data. The training data sources are shown in Table 2 . The published dates of the above sources range from Japan-ruled era to postwar era (1945-) . Two paragraphs are selected from each book, with a total of 614 syllables (438 word tokens).", | |
| "cite_spans": [ | |
| { | |
| "start": 182, | |
| "end": 189, | |
| "text": "(1945-)", | |
| "ref_id": null | |
| } | |
| ], | |
| "ref_spans": [ | |
| { | |
| "start": 89, | |
| "end": 96, | |
| "text": "Table 2", | |
| "ref_id": "TABREF2" | |
| } | |
| ], | |
| "eq_spans": [], | |
| "section": "Data", | |
| "sec_num": "2.1" | |
| }, | |
| { | |
| "text": "In addition to data drawn from the same project, the test data also include some other sources we collected. Four sources are selected as well. The test data sources are shown in Table 3 . Two or three paragraphs are selected from each book or article. The test data total 962 syllables (656 word tokens) and also cover two eras but with a longer time span.", | |
| "cite_spans": [], | |
| "ref_spans": [ | |
| { | |
| "start": 179, | |
| "end": 186, | |
| "text": "Table 3", | |
| "ref_id": "TABREF3" | |
| } | |
| ], | |
| "eq_spans": [], | |
| "section": "Data", | |
| "sec_num": "2.1" | |
| }, | |
| { | |
| "text": "As there is no standard on part of speech (POS) for Taiwanese at present, we use the standard of Chinese instead (see Results section). We obtain the corresponding Chinese translation for each Taiwanese word by looking up the Taiwanese-Chinese On-line Dictionary. [Iunn 2003 ] We, then, look up the POS of the Chinese in the 80,000-word CED. Ambiguity encountered includes:", | |
| "cite_spans": [ | |
| { | |
| "start": 264, | |
| "end": 274, | |
| "text": "[Iunn 2003", | |
| "ref_id": "BIBREF3" | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Part of Speech Tagging", | |
| "sec_num": "2.2" | |
| }, | |
| { | |
| "text": "(a) homonymy, especially monosyllabic homonyms;", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Part of Speech Tagging", | |
| "sec_num": "2.2" | |
| }, | |
| { | |
| "text": "(b) one-to-many mapping when mapping Taiwanese to Chinese;", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Part of Speech Tagging", | |
| "sec_num": "2.2" | |
| }, | |
| { | |
| "text": "(c) multiple possible POSs for each Chinese word.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Part of Speech Tagging", | |
| "sec_num": "2.2" | |
| }, | |
| { | |
| "text": "To resolve homonymy, we choose the word with the highest querying frequency. We found out that this strategy works under most situations. Due to the fact that one Taiwanese word may map to multiple Chinese words, and one Chinese word could possibly have multiple POSs, there may be multiple POSs for one Taiwanese word. We initially retain all candidate POSs in tagging and only attempt to narrow down the list upon applying the sandhi algorithm. Of the 46 POSs in the Chinese Electronic Dictionary, we adopt the top level and adjust certain POSs known to affect tone sandhi. For example, Vh (state intransitive verb, etc.) is marked A, Nh (pronoun) marked R, Ng (postposition) marked G, and Nd (time) marked S. The POS classes we used are shown in Table 4 . As for unknown words, if they are of the form 'XX' or 'XXX' (duplicate or triplicate syllables), we mark them as A (adjective). Other words are marked as N (noun).", | |
| "cite_spans": [], | |
| "ref_spans": [ | |
| { | |
| "start": 749, | |
| "end": 756, | |
| "text": "Table 4", | |
| "ref_id": "TABREF4" | |
| } | |
| ], | |
| "eq_spans": [], | |
| "section": "Part of Speech Tagging", | |
| "sec_num": "2.2" | |
| }, | |
| { | |
| "text": "The marks representing tone sandhis are listed in Table 5 . Words with normal sandhi are usually not marked. ", | |
| "cite_spans": [], | |
| "ref_spans": [ | |
| { | |
| "start": 50, | |
| "end": 57, | |
| "text": "Table 5", | |
| "ref_id": "TABREF5" | |
| } | |
| ], | |
| "eq_spans": [], | |
| "section": "Tone Sandhi Marks", | |
| "sec_num": "2.3" | |
| }, | |
| { | |
| "text": "Tone sandhi rules are the most important part of this study. The algorithm for sandhi marking is shown in Table 6 . These sandhi rules work on 4 different levels: the syllable, the word, the part of speech, and the sentence pattern.", | |
| "cite_spans": [], | |
| "ref_spans": [ | |
| { | |
| "start": 106, | |
| "end": 113, | |
| "text": "Table 6", | |
| "ref_id": "TABREF6" | |
| } | |
| ], | |
| "eq_spans": [], | |
| "section": "Tone Sandhi Rules", | |
| "sec_num": "2.4" | |
| }, | |
| { | |
| "text": "The algorithm described above is mainly based on a) tone sandhi rules proposed by linguists; b) rules induced from the training data; and c) our intuition as native-speaking observers of sandhi phenomena. We also consulted d) the word segmentation results of the CKIP (examining its POS tagging output) and e) the Taiwanese concordancer system (to check the sandhi phenomena of certain words) when we met some questions.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Tone Sandhi Rules", | |
| "sec_num": "2.4" | |
| }, | |
| { | |
| "text": "It should be noted that some of the sandhi rules proposed by linguists deal with specific contexts and thus cannot be broadly applied; some others carry exceptions. There is, therefore, some difficulty in converting these rules into an algorithm. So, besides (a), we also formulated some rules from (b) and (c) by analyzing errors in the training data output. In principle sandhi rules are formulated to be applicable to \"most situations\" --i.e. an accuracy rate of over 75% on corpus data. Once applied, the new rules may affect the original rules, so (d) and (e) are our important references in deciding whether or not to apply the new rules. the last b\u0113 is marked as basic tone.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Tone Sandhi Rules", | |
| "sec_num": "2.4" | |
| }, | |
| { | |
| "text": "Moreover, because of the uncertainty in tagging POS, some rules are set to apply only when there is no ambiguity, while some other rules are applied to any matching POSs.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Tone Sandhi Rules", | |
| "sec_num": "2.4" | |
| }, | |
| { | |
| "text": "We currently employ 20 rules and expect to refine them or append new ones.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Tone Sandhi Rules", | |
| "sec_num": "2.4" | |
| }, | |
| { | |
| "text": "The following training data represents a pre-tagged source (Chinese and English translations added): After POS tagging and applying the sandhi rules: The letters within the parentheses are the POSs. Incorrectly processed syllables are boxed.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Tone Sandhi Rules", | |
| "sec_num": "2.4" | |
| }, | |
| { | |
| "text": "(10) Chhin-chhi\u016b\u207f [ \u50cf ] \u00e1n-ni[ \u9019 \u6a23 ] l\u00e2i[ \uf92d ] k\u00f3ng[\uf96f] , ch\u0101i[\u5728] l\u00e1n[\u6211\u5011] T\u00e2i-\u00f4an[\u53f0\u7063] k\u012bn-k\u012bn[ \u8fd1 \u8fd1 ] chi\u030dt-tiap-\u00e1-k\u00fa[ \u4e00 \u4e0b \u5b50 ] \u00ea[ \u7684 ] kang-hu[\u5de5\u592b] , \u00e0i[\u8981] soa\u207f[\u5c71] chi\u016b[\u5c31] \u016b[\u6709] soa\u207f[ \u5c71 ] , \u00e0i[ \u8981 ] h\u00e1i[ \u6d77 ] chi\u016b[ \u5c31 ] \u016b[ \u6709 ] h\u00e1i[ \u6d77 ] ,", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Tone Sandhi Rules", | |
| "sec_num": "2.4" | |
| }, | |
| { | |
| "text": "(11) Chhin -chhi\u016b\u207f(D) \u00e1n-ni#(D;N) l\u00e2i(D;V) k\u00f3ng#(V), ch\u0101i(D;A;P;V) l\u00e1n(R) T\u00e2i-\u00f4an#(N) k\u012bn-k\u012bn(A) chi\u030dt-tiap&-\u00e1-k\u00fa#(N) \u00ea(M) kang-hu#(A;N), \u00e0i(D;V) soa\u207f# (N) chi\u016b(D) \u016b(D;P;V) soa\u207f#(N), \u00e0i(D;V) h\u00e1i#(N) chi\u016b(D) \u016b(D;P;V) h\u00e1i#(N), beh$(D) jo\u030dah#(A) chi\u016b(D) \u016b(D;P;V) jo\u030dah#(A), k\u00f4a\u207f#(A) chi\u016b(D) \u016b(D;P;V) k\u00f4a\u207f#(A). S\ue023-\u00ed(C) thang(D) k\u00f3ng(V) T\u00e2i-\u00f4an#(N) s\u012b(D;V) chi\u030dt-\u00ea#(N) si\u00f3(D;A) Tang-i\u00fb\u207f#(N). L\u00e1n(R)", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Tone Sandhi Rules", | |
| "sec_num": "2.4" | |
| }, | |
| { | |
| "text": "Three authors of this paper, who are skilled native speakers familiar with written Taiwanese, evaluated the correctness of the output. Note that in certain contexts more than one sandhi result is acceptable, and depending on discourse considerations some speakers may opt for one sandhi result over others. For example, \"h\ue029 l\u00ed\" [\u7d66\u4f60,\"give you\"] can be read as (3,2) (normal sandhi) or (7,7) (following sandhi). Telephone number is another example: the number may be divided into various groups, each group containing 2, 3 or 4 digits.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Evaluation", | |
| "sec_num": "3.1" | |
| }, | |
| { | |
| "text": "There are 614 syllables of training data, 16 errors, giving an accuracy rate of 97.39%. There are 962 syllables of test data with 106 errors, or an accuracy rate of 88.98%. Table 7 shows the number of errors and accuracy rate for each paragraph. Table 8 shows the numbers of each rule applied in training data and test data respectively. We count the number of affected syllables, accurately affected syllables, and accuracy rate of each rule. Note that rules 5 & 6 don't seem work well because of POS ambiguities, rule 7 does not affect any syllables because the word whose POS is G (postposition) also has other POSs, rule 14 does not affect any syllables because there are no triplicated words in our training and test data. Table 9 shows the number of dominant rule, accurate dominant rule, and accuracy rate. After examination, we find that we can add 7 additional rules without too much effort; in this way, we were able to fix 20 errors and achieve a 91.06% accuracy rate. Table 10 shows the additional rules in order to fix 20 errors in test data.", | |
| "cite_spans": [], | |
| "ref_spans": [ | |
| { | |
| "start": 173, | |
| "end": 180, | |
| "text": "Table 7", | |
| "ref_id": "TABREF9" | |
| }, | |
| { | |
| "start": 246, | |
| "end": 253, | |
| "text": "Table 8", | |
| "ref_id": "TABREF10" | |
| }, | |
| { | |
| "start": 728, | |
| "end": 736, | |
| "text": "Table 9", | |
| "ref_id": "TABREF11" | |
| }, | |
| { | |
| "start": 981, | |
| "end": 989, | |
| "text": "Table 10", | |
| "ref_id": "TABREF1" | |
| } | |
| ], | |
| "eq_spans": [], | |
| "section": "Preliminary Results", | |
| "sec_num": "3.2" | |
| }, | |
| { | |
| "text": "test data Word suffix \"V-tit\" (adverbialize the word whose POS is verb) 5 Double sandhi of \"khah\" [\uf901, \"more\"] 4", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Number of corrections in", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "Re-process the syllable preceding \"\u00ea\" [\u500b, a numerary adjunct] when the preceding word is a number or \"chit/hit/pa\u030dt\" [\u9019/\u90a3/\u5225, \"this/that/other\"] 4 \"V-ji\u030dp-l\u00e2i\" [V \u9032\uf92d, \"V-in\"]: mark as neutral sandhi when sentence-final 3", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Number of corrections in", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "Word \"hut-ji\u00e2n\" [\u5ffd\u7136, \"suddenly\"]: mark the last syllable as basic tone in any case 2", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Number of corrections in", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "Word \"k\u012bn-l\u00e2i\" [\u8fd1\uf92d, \"recently\"]: mark the last syllable as basic tone in any case 1", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Number of corrections in", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "Word suffix \"N-nih\"[N \uf9e8, \"inside N\"]: mark as neutral sandhi 1", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Number of corrections in", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "Some of the problems we encountered may be taken into account in the future.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Analysis of Errors and Relevant Issues", | |
| "sec_num": "4." | |
| }, | |
| { | |
| "text": "In our investigation, we use the POS set for Chinese. Whether this approach is suitable for Taiwanese is a debatable linguistic question requiring further investigation. Although a few studies of the POS of Taiwanese are available from as early as the 1930s, currently these data have yet to be digitized, and will need to be reviewed by linguists to ensure that they are suitable for dealing with the sandhi problem. [Tseng 1997 ] proposes a standard for Taiwanese word segmentation. Unfortunately discussion is lagging. Should a working word segmentation standard emerge, we would also need a dictionary conforming to that standard.", | |
| "cite_spans": [ | |
| { | |
| "start": 418, | |
| "end": 429, | |
| "text": "[Tseng 1997", | |
| "ref_id": "BIBREF10" | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "POS", | |
| "sec_num": "4.1" | |
| }, | |
| { | |
| "text": "Historically, the use of Han script to represent Taiwanese has suffered from a high degree of idiosyncrasy in character choice. For documents written in Latin script, most of the differences attributed to dialects can be reconciled by referencing existing dictionaries. Orthographic inconsistency in the use of hyphen is more problematic, as it could affect the result of sandhi processing. Manual standardization of hyphen placement is hardly a solution.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Standardization of Written Taiwanese", | |
| "sec_num": "4.3" | |
| }, | |
| { | |
| "text": "We have encountered certain sandhi problems that likely cannot be solved solely by inspecting the POS order. These include verb-verb (VV) and noun-noun (NN) patterns: (12) is an example of a VV pattern. The final syllable of the first verb in (a) should be marked as sandhi tone, while in (b) it should be marked as basic tone. Differences in the internal structure of these two initial verbs suggest some clues for handling this problem. However, its implementation awaits further research. (13) a. \"ti\u0101n-s\u012b k\u00f3ng-k\u00f2\" [\u96fb\u8996(\u7684)\u5ee3\u544a, \"TV advertisement\"] b. \"th\u00e2ng-th\u014da chi\u00e1u-chiah\" [\u6606\u87f2(\u3001)\u5c0f\u9ce5, \"insects and birds\"] (13) is an example of a NN pattern. Again, the final syllable of the first noun in (a) should be marked as sandhi tone, while in (b) it should be marked as basic tone. Currently, we see no solution to this.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Tone Sandhi Problems Not Solvable by POS Order", | |
| "sec_num": "4.4" | |
| }, | |
| { | |
| "text": "Error conditions, including those discussed in the previous sections, are listed below with possible in ", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Error Conditions", | |
| "sec_num": "4.5" | |
| }, | |
| { | |
| "text": "A three-year-old child native speaker can process tone sandhi correctly and apparently without effort, yet it is rather more difficult for a computer system to do so. Clearly, a practical system for sandhi processing of Taiwanese remains out-of-reach and a cause for future research. Some suggestions for future work:", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Future Work", | |
| "sec_num": "5." | |
| }, | |
| { | |
| "text": "(a) Solicit assistance from linguists. It is hoped that linguistics will define a standard for part-of-speech analysis and word segmentation, and that a dictionary conforming to such a standard will be built.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Future Work", | |
| "sec_num": "5." | |
| }, | |
| { | |
| "text": "(b) Improve word segmentation, especially the processing of morphology, quantitative words, and proper nouns.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Future Work", | |
| "sec_num": "5." | |
| }, | |
| { | |
| "text": "(c) Improve the processing of POS tags to account for ambiguity.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Future Work", | |
| "sec_num": "5." | |
| }, | |
| { | |
| "text": "(d) Improve the dictionary's part-of-speech data, such as making use of Embree's POS analysis [Embree 1984] .", | |
| "cite_spans": [ | |
| { | |
| "start": 94, | |
| "end": 107, | |
| "text": "[Embree 1984]", | |
| "ref_id": "BIBREF2" | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Future Work", | |
| "sec_num": "5." | |
| }, | |
| { | |
| "text": "(e) Improve the sandhi rules.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Future Work", | |
| "sec_num": "5." | |
| }, | |
| { | |
| "text": "(f) Find alternative ways of modeling sandhi processing, such as Cheng's grammar template model. [Cheng 2002] ", | |
| "cite_spans": [ | |
| { | |
| "start": 97, | |
| "end": 109, | |
| "text": "[Cheng 2002]", | |
| "ref_id": "BIBREF1" | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Future Work", | |
| "sec_num": "5." | |
| } | |
| ], | |
| "back_matter": [ | |
| { | |
| "text": "This work is supported by National Museum of Taiwanese Literature in Taiwan. We also thank the anonymous reviewers for their constructive opinions.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Acknowledgements", | |
| "sec_num": null | |
| } | |
| ], | |
| "bib_entries": { | |
| "BIBREF0": { | |
| "ref_id": "b0", | |
| "title": "Taiwanese and Mandarin Structures and Their Developmental Trends in Taiwan Book I : Taiwanese Phonology and Morphology", | |
| "authors": [ | |
| { | |
| "first": "R", | |
| "middle": [], | |
| "last": "Cheng", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1997, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Cheng, R., Taiwanese and Mandarin Structures and Their Developmental Trends in Taiwan Book I : Taiwanese Phonology and Morphology, Yuan-liou Publishing Co., 1997.", | |
| "links": null | |
| }, | |
| "BIBREF1": { | |
| "ref_id": "b1", | |
| "title": "Tone Sandhi on the Grammar Template--Cognition and Testing", | |
| "authors": [ | |
| { | |
| "first": "R", | |
| "middle": [], | |
| "last": "Cheng", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2002, | |
| "venue": "Proceeding of 2002 International Conference on Teaching and Researching of Taiwanese Romanization", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "1--19", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Cheng, R., \"Tone Sandhi on the Grammar Template--Cognition and Testing,\" Proceeding of 2002 International Conference on Teaching and Researching of Taiwanese Romanization, 2002, pp. I1-I19.", | |
| "links": null | |
| }, | |
| "BIBREF2": { | |
| "ref_id": "b2", | |
| "title": "A Dictionary of Southern Min", | |
| "authors": [ | |
| { | |
| "first": "B", | |
| "middle": [ | |
| "L M A" | |
| ], | |
| "last": "Embree", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1984, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Embree, B. L.M.A., A Dictionary of Southern Min. Taipei Language Institute, 1984.", | |
| "links": null | |
| }, | |
| "BIBREF3": { | |
| "ref_id": "b3", | |
| "title": "Taiwanese-Chinese On-line Dictionary --Discussion of Building Technique and its Utilization", | |
| "authors": [ | |
| { | |
| "first": "U.-G", | |
| "middle": [], | |
| "last": "Iunn", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2003, | |
| "venue": "Proceeding of 3 rd International Conference on Internet Chinese Education", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "132--141", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Iunn, U.-G., \"Taiwanese-Chinese On-line Dictionary --Discussion of Building Technique and its Utilization,\" Proceeding of 3 rd International Conference on Internet Chinese Education, 2003, pp. 132-141.", | |
| "links": null | |
| }, | |
| "BIBREF4": { | |
| "ref_id": "b4", | |
| "title": "Review and Analysis of Taiwan Ho-lo Language non-Han Character Spelling Symbols", | |
| "authors": [ | |
| { | |
| "first": "U.-G", | |
| "middle": [], | |
| "last": "Iunn", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "H.-K", | |
| "middle": [], | |
| "last": "Tiunn", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1999, | |
| "venue": "Proceedings of 1 st Conference on the Regeneration and Rebuild of Taiwan Mother Tongue Culture", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "62--76", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Iunn, U.-G. and H.-K. Tiunn, \"Review and Analysis of Taiwan Ho-lo Language non-Han Character Spelling Symbols,\" Proceedings of 1 st Conference on the Regeneration and Rebuild of Taiwan Mother Tongue Culture, 1999, pp. 62-76.", | |
| "links": null | |
| }, | |
| "BIBREF5": { | |
| "ref_id": "b5", | |
| "title": "A Survey of Media and Data Processing Development for Written Taiwanese", | |
| "authors": [ | |
| { | |
| "first": "U.-G", | |
| "middle": [], | |
| "last": "Iunn", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "H", | |
| "middle": [ | |
| "H" | |
| ], | |
| "last": "Tan-Tenn", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": null, | |
| "venue": "Accepted by International Journal of the Sociology of Language", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Iunn, U.-G. and H. H.Tan-Tenn, \"A Survey of Media and Data Processing Development for Written Taiwanese,\" Accepted by International Journal of the Sociology of Language, Special Issues on Taiwanese.", | |
| "links": null | |
| }, | |
| "BIBREF6": { | |
| "ref_id": "b6", | |
| "title": "A Taiwanese Text-to-Speech System with Applications to Language Learning", | |
| "authors": [ | |
| { | |
| "first": "M.-S", | |
| "middle": [], | |
| "last": "Liang", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "J.-C", | |
| "middle": [], | |
| "last": "Yang", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Y.-C", | |
| "middle": [], | |
| "last": "Chiang", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "R.-Y", | |
| "middle": [], | |
| "last": "Lyu", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2004, | |
| "venue": "Proceedings of the 4 th IEEE International Conference on Advanced Learning Technologies", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "91--95", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Liang, M.-S., J.-C. Yang, Y.-C. Chiang, and R.-Y. Lyu, \"A Taiwanese Text-to-Speech System with Applications to Language Learning,\" Proceedings of the 4 th IEEE International Conference on Advanced Learning Technologies, 2004, pp. 91-95.", | |
| "links": null | |
| }, | |
| "BIBREF7": { | |
| "ref_id": "b7", | |
| "title": "A Mandarin to Taiwanese Min Nan Machine Translation System with Speech Synthesis of Taiwanese Min Nan", | |
| "authors": [ | |
| { | |
| "first": "C.-J", | |
| "middle": [], | |
| "last": "Lin", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "H.-H", | |
| "middle": [], | |
| "last": "Chen", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1999, | |
| "venue": "International Journal of Computational Linguistics and Chinese Language Processing", | |
| "volume": "4", | |
| "issue": "1", | |
| "pages": "59--84", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Lin, C.-J. and H.-H. Chen, \"A Mandarin to Taiwanese Min Nan Machine Translation System with Speech Synthesis of Taiwanese Min Nan,\" International Journal of Computational Linguistics and Chinese Language Processing, 4(1), 1999, pp. 59-84.", | |
| "links": null | |
| }, | |
| "BIBREF8": { | |
| "ref_id": "b8", | |
| "title": "The Study of Minnan Vocabulary in Taiwan", | |
| "authors": [ | |
| { | |
| "first": "G.-C", | |
| "middle": [], | |
| "last": "Lu", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1999, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Lu, G.-C., The Study of Minnan Vocabulary in Taiwan. SMC Publishing Inc., 1999.", | |
| "links": null | |
| }, | |
| "BIBREF9": { | |
| "ref_id": "b9", | |
| "title": "Principles of POJ or the Taiwanese Orthography: An Introduction to Its Sound-Symbol Correspondences and Related Issues", | |
| "authors": [ | |
| { | |
| "first": "J.-H", | |
| "middle": [], | |
| "last": "Tiunn", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Y.-H", | |
| "middle": [], | |
| "last": "Chang", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2001, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Tiunn, J.-H. (Chang, Y.-H.). Principles of POJ or the Taiwanese Orthography: An Introduction to Its Sound-Symbol Correspondences and Related Issues, Crane Publishing Co., 2001.", | |
| "links": null | |
| }, | |
| "BIBREF10": { | |
| "ref_id": "b10", | |
| "title": "The Project Report for the Collecting, Cataloging and Select Editing of Taiwanese Liturature Publications", | |
| "authors": [ | |
| { | |
| "first": "C.-C", | |
| "middle": [], | |
| "last": "Tseng", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1997, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "47--73", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Tseng, C.-C., \"The Discussion of Taiwanese Word Segmentation Principles,\" The Project Report for the Collecting, Cataloging and Select Editing of Taiwanese Liturature Publications, pp. 47-73, Council for Culture Affairs, 1997.", | |
| "links": null | |
| }, | |
| "BIBREF11": { | |
| "ref_id": "b11", | |
| "title": "Online Resources Chinese On-line Word Segmentation System", | |
| "authors": [], | |
| "year": null, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Online Resources Chinese On-line Word Segmentation System, http://ckipsvr.iis.sinica.edu.tw", | |
| "links": null | |
| }, | |
| "BIBREF12": { | |
| "ref_id": "b12", | |
| "title": "Digital Archive Database for Written Taiwanese", | |
| "authors": [], | |
| "year": null, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Digital Archive Database for Written Taiwanese, http://iug.csie.dahan.edu.tw/nmtl/dadwt", | |
| "links": null | |
| } | |
| }, | |
| "ref_entries": { | |
| "FIGREF0": { | |
| "num": null, | |
| "text": "Taiwanese Southern-Min tone sandhi system diagram", | |
| "type_str": "figure", | |
| "uris": null | |
| }, | |
| "FIGREF1": { | |
| "num": null, | |
| "text": "Mark the last syllable as basic tone # 3 \u00ea [\u7684, \"of\"] : Mark the syllable preceding \u00ea as basic tone # \u00ea is a special marker 4 A/A Pair 4.1 A/A Pair: Mark the last syllable of the first word as basic tone # POS level, with ambiguity 5 N/V, N/A, N/P, N/R, and N/D Pairs 5.1 N/V Pair: Mark the last syllable of the first word as basic tone # 5.2 N/A Pair: Mark the last syllable of the first word as basic tone # 5.3 N/P Pair: Mark the last syllable of the first word as basic tone # 5.4 N/R Pair: Mark the last syllable of the first word as basic tone # 5.5 N/D Pair: Mark the last syllable of the first word as basic tone # POS level , with ambiguity 6 C: Mark the last syllable of the preceding word as basic tone # POS level 7 G: Mark the last syllables of both the preceding word and the word itself as basic tones #i / in [\u4ed6(\u5011), \"(s)he/they\"] : Mark them as normal sandhi even if they are the last syllables 9.2 g\u00f3a / l\u00ed / g\u00fan / g\u00f3an / l\u00e1n / l\u00edn [\u6211/\u4f60(\u5011)(\u7684), \"I/you/my/our/your\"]of POS R: Mark them as normal sandhi if they are not the last syllable POS/Word level 10 Sentence-final k\u00f3ng [\u8b1b, \"say\"] : Mark this word as normal sandhi if the delimiter is among [,\uff1a: \"] and there is any word of POS R in front of this word (note: this rule needs to be refined in case there is a name in front of this word) Word level, induced from training data", | |
| "type_str": "figure", | |
| "uris": null | |
| }, | |
| "FIGREF2": { | |
| "num": null, | |
| "text": "Some rules have priority. Subsequent rules can supersede previous ones. As an example, rule 9 (pronoun rule) can supersede rule 3 (of rule). At the level of sentence pattern, rule 19.4.2 can supersede 19.4.1 as in the following example: (9) \"L\u00ed \u0113 kh\u00ec kok-g\u014da b\u0113\" [\u4f60\u6703\uf967\u6703\u53bb\u570b\u5916? \"Will you go abroad or not\"]:the last b\u0113 [\uf967\u6703, \"will not\"] is marked as neutral sandhi, whereas \"L\u00ed \u0113 kh\u00ec kok-g\u014da iah-s\u012b b\u0113\" [\u4f60\u6703\uf967\u6703\u53bb\u570b\u5916? \"Will you go abroad or not\"]:", | |
| "type_str": "figure", | |
| "uris": null | |
| }, | |
| "FIGREF3": { | |
| "num": null, | |
| "text": "T\u00e2i-\u00f4an#(N) \u016b(D;P;V) chit-kh\u00f3an#(D;N) thian-ji\u00e2n#(A) \u00ea(M) h\u00f3-k\u00e9ng#(N), h\u00f3(D;A;C;V) kh\u00ec-h\u0101u#(N), chiong-l\u00e2i#(S) n\u0101-s\u012b(C) \u0113ng-sim#(N) ke(V) l\u00e2ng#(N) \u00ea(M) kang-hu#(A;N) t\u014da-t\u014da(A) l\u00e2i(D;V) ch\u00e9ng-t\u00f9n#(V), tek-khak(D) \u0113(D;V) chi\u00e2\u207f-ch\u00f2(V)Tang-i\u00fb\u207f#(N) \u00ea(M) t\u014da(A;N) kong-hn\u0302g#(N), h\ue029(D;P;V) Tang-i\u00fb\u207f#(N) \u00ea(M) l\u00e2ng#(N) chi\u030dp-\u00f3a(V) l\u00e2i(D;V) hi\u00f3ng-hok#(A) an-lo\u030dk#(A).", | |
| "type_str": "figure", | |
| "uris": null | |
| }, | |
| "FIGREF4": { | |
| "num": null, | |
| "text": "12) a. \"phah-pi\u00e0\u207f(V) ch\u00f2(V) khang-kh\u00f2e(kh\u00e8) (N)\" (2,2,2,7,3) [\u52aa\uf98a\u505a\u5de5\u4f5c, \"do work hard\"] b. \"kia\u030dh-ba\u030dk(V) kh\u00f2a\u207f(V) hn\u0302g(N)\" (3,8,2,5) [\u8209\u76ee\u770b\u5712, \"lift eyes and seeplowland\"]", | |
| "type_str": "figure", | |
| "uris": null | |
| }, | |
| "TABREF1": { | |
| "text": "", | |
| "content": "<table><tr><td>Normal</td><td>Basic tone of syllable</td><td>1</td><td>2</td><td>3</td><td>4</td><td>5</td><td/><td>8</td></tr><tr><td colspan=\"2\">sandhi Sandhi tone</td><td>7</td><td>1</td><td>2</td><td colspan=\"2\">8/2 7/3</td><td/><td>4/3</td></tr><tr><td>Following</td><td>Basic tone of preceding syllable</td><td>1</td><td>2</td><td>3</td><td>4</td><td>5</td><td/><td>8</td></tr><tr><td colspan=\"2\">sandhi Sandhi tone</td><td>1</td><td>3</td><td>3</td><td>3</td><td>7</td><td/><td>1</td></tr><tr><td>Neutral</td><td>Basic tone of preceding syllable</td><td>1</td><td>2</td><td>3</td><td>4</td><td>5</td><td/><td>8</td></tr><tr><td colspan=\"2\">sandhi Sandhi tone</td><td>3</td><td>3</td><td>3</td><td>4/3</td><td>3</td><td/><td>4/3</td></tr><tr><td>Double</td><td>Basic tone of syllable</td><td>-</td><td>-</td><td>3</td><td>4</td><td>-</td><td>-</td><td>-</td></tr><tr><td colspan=\"2\">sandhi Sandhi tone</td><td/><td/><td>1</td><td>1</td><td/><td/><td/></tr><tr><td>Pre-\u00e1</td><td>Basic tone of preceding syllable of \"\u00e1\"</td><td>1</td><td>2</td><td>3</td><td>4</td><td>5</td><td>7</td><td>8</td></tr><tr><td colspan=\"2\">sandhi Sandhi tone</td><td>7</td><td>1</td><td>1</td><td>8/1</td><td>7</td><td/><td>4/7</td></tr><tr><td>Triplicate</td><td>Basic tone of the first syllable of three</td><td>1</td><td>2</td><td>3</td><td>4</td><td>5</td><td/><td>8</td></tr><tr><td colspan=\"2\">sandhi Sandhi tone</td><td>5</td><td>1</td><td>2</td><td>8/2</td><td>5</td><td/><td>5</td></tr></table>", | |
| "html": null, | |
| "num": null, | |
| "type_str": "table" | |
| }, | |
| "TABREF2": { | |
| "text": "", | |
| "content": "<table><tr><td>Book or aticle</td><td>year</td><td>author</td><td>genre</td></tr><tr><td>\"Sin-b\u00fbn \u00ea cha\u030dp-lio\u030dk\" [\u65b0\u805e\u7684\u96dc\uf93f, \"News Bulletin\"]</td><td>1913</td><td>unknown</td><td>journalism</td></tr><tr><td>\"Cha\u030dp-h\u0101ng k\u00f3an-ki\u00e0n\" [\u5341\u9805\u7ba1\ufa0a, \"Ten Humble Opinions\"]</td><td/><td>Chh\u00f2a P\u00f4e-h\u00f3e [\u8521\u57f9\u706b]</td><td>discourse</td></tr><tr><td>\"Chh\u00e1u-tui t\u00e9ng \u00ea b\u00een-b\u0101ng --j\u00ee-t\u00f4ng chong-k\u00e0u k\ue025-s\u016b\" [\u8349\u5806\u4e0a\u7684\u5922\u2500\u5152\u7ae5\u5b97\u6559\u6545\u4e8b, \"Dreams on the Grass Stack --Religious Stories for Children\"]</td><td>1955</td><td>N\u0302g H\u00f4ai-un [\u9ec3\u61f7\u6069]</td><td>short stories</td></tr><tr><td>\"Tang-p\ue029 th\u00f4an-t\u014d ki\u00e0n-b\u00fbn k\u00ec\" [\u6771\u90e8\u50b3\u9053\ufa0a\u805e\u8a18, \"Record of Preaching in Eastern Taiwan\"]</td><td>1961</td><td>T\u00e2n K\u00e0ng-h\u00e2ng [\u9673\ufa09\u7965]</td><td>journalism</td></tr></table>", | |
| "html": null, | |
| "num": null, | |
| "type_str": "table" | |
| }, | |
| "TABREF3": { | |
| "text": "", | |
| "content": "<table><tr><td>Book or article</td><td>year</td><td>author</td><td>genre</td></tr><tr><td>\"Pe\u030dh-\u014de-j\u012b \u00ea l\u012b-ek\" [\u767d\u8a71\u5b57\u7684\uf9dd\u76ca, \"The Benefits of Using Peh-oe-ji\"]</td><td colspan=\"2\">1885 Reverend Ia\u030dp [\uf96e\u7267\u5e2b]</td><td>discourse</td></tr><tr><td>\"Kau-chi\u00e0n \u00ea Siau-sit\" [\u4ea4\u6230\u7684\u6d88\u606f, \"News of the War\"]</td><td>1905</td><td>the editorial office of T\u00e2i-l\u00e2m Prefectural Church News</td><td>report</td></tr><tr><td>\"Thi\u00e0\u207f l\u00ed i\u00e2\u207f k\u00e8 thong s\u00e8-kan\" [\u75bc\u611b\u4f60\u52dd\u904e\u5168</td><td/><td/><td/></tr><tr><td>\u4e16\u754c, \"Caring About You More Than the</td><td colspan=\"2\">1955 L\u014da J\u00een-seng [\u8cf4\u4ec1\u8072]</td><td>novel</td></tr><tr><td>Whole World\"]</td><td/><td/><td/></tr><tr><td>\"\u00c0i l\u00ed kap \u00e0i i p\u00ee\u207f-\u00e1 ch\u014de\" [\u611b\u59b3\u548c\u5979\u4e00\u6a23\u591a, \"Loving You as Much as Her\"]</td><td colspan=\"2\">1997 L\ue027 T\u00e0n-chhun [\uf933\u8a95\u6625]</td><td>prose</td></tr></table>", | |
| "html": null, | |
| "num": null, | |
| "type_str": "table" | |
| }, | |
| "TABREF4": { | |
| "text": "", | |
| "content": "<table><tr><td>POS</td><td>statement</td><td>POS</td><td>statement</td><td>POS</td><td>statement</td></tr><tr><td>A</td><td>adjective</td><td>I</td><td>interjection</td><td>R</td><td>pronoun</td></tr><tr><td>C</td><td>conjunction</td><td>M</td><td>special marker</td><td>S</td><td>time</td></tr><tr><td>D</td><td>adverb</td><td>N</td><td>noun</td><td>T</td><td>auxiliary</td></tr><tr><td>G</td><td>postposition</td><td>P</td><td>preposition</td><td>V</td><td>verb</td></tr></table>", | |
| "html": null, | |
| "num": null, | |
| "type_str": "table" | |
| }, | |
| "TABREF5": { | |
| "text": "", | |
| "content": "<table><tr><td>Symbol</td><td>Phenomenon</td><td>Symbol</td><td>Phenomenon</td></tr><tr><td>(none)</td><td>Normal sandhi</td><td>$</td><td>Double sandhi</td></tr><tr><td>#</td><td>Basic tone</td><td>&</td><td>Pre-\u00e1 sandhi</td></tr><tr><td>@</td><td>Following sandhi</td><td>~</td><td>Triplicate sandhi</td></tr><tr><td>%</td><td>Neutral sandhi</td><td>^</td><td>Rising sandhi</td></tr></table>", | |
| "html": null, | |
| "num": null, | |
| "type_str": "table" | |
| }, | |
| "TABREF6": { | |
| "text": "", | |
| "content": "<table><tr><td/><td>Rule</td><td>Remark</td></tr><tr><td>1</td><td>Apply normal sandhi to all syllables</td></tr></table>", | |
| "html": null, | |
| "num": null, | |
| "type_str": "table" | |
| }, | |
| "TABREF7": { | |
| "text": "", | |
| "content": "<table><tr><td>11 pre-\u00e1 [\u00e1 is suffixof a word]: Mark any syllables just before \u00e1 as pre-\u00e1</td><td>Syllable level</td></tr><tr><td>sandhi &</td><td/></tr><tr><td>12 Double sandhi</td><td/></tr><tr><td>12.1 beh [\u8981, \"want\"] : Mark any beh as double sandhi $ unless it</td><td>Syllable level</td></tr><tr><td>appears at the end, including those within a word, such as</td><td/></tr><tr><td>kiong-beh, tih-beh .</td><td/></tr><tr><td>12.2 kh\u00ec [\u53bb, \"go\"] : Mark kh\u00ec as double sandhi $ if the POS of the</td><td>Word level</td></tr><tr><td>immediately following word is N or V, unless it appears at the</td><td/></tr><tr><td>end</td><td/></tr><tr><td>12.3 koh [\u518d, \"again\"] : Mark any koh as double sandhi $, including</td><td>Syllable level,</td></tr><tr><td>those within a word, such as chiah-koh[\u518d, \"and then\"] or</td><td>extended from</td></tr><tr><td>i\u00e1u-koh[\u9084\u662f, \"still\"], unless it appears at the end</td><td>training data</td></tr><tr><td>12.4 kah[\u548c, \"and\"] : Mark any kah as double sandhi $ unless it</td><td>Word level</td></tr><tr><td>appears at the end</td><td/></tr><tr><td>13 Neutral sandhi of --: Mark the syllable just before --as basic tone,</td><td>Word level</td></tr><tr><td>and mark each syllable after --as neutral sandhi %</td><td/></tr><tr><td>14 Triplicate sandhi: Mark the first syllable as triplicate sandhi if that</td><td>Word level</td></tr><tr><td>word has 3 syllables of the same spelling</td><td/></tr><tr><td>15 Special words</td><td>Word level, extend</td></tr><tr><td>15.1 s\u00edm-mih / s\u00edm-mi\u030dh[\uf9fd\u9ebc, \"what\"] : Change these words into s\u00edm-m\u00ed (sandhi marks not changed) 15.2 \u00e1n-ni / \u00e0n-ni / an-ni / an-n\u012b [\u9019\u6a23, \"thus\"] : Change these words</td><td>from training data beacuse of not yet standardized</td></tr><tr><td>into \u00e1n-ni and to mark its sandhi marks as t#</td><td/></tr><tr><td>16 Markers</td><td/></tr><tr><td>16.1 iah-s\u012b / ah-s\u012b / i\u00e1h-s\u012b / \u00e1h-s\u012b / \u00e1-s\u012b [\u6216\u662f, \"or\"] : Mark the last</td><td>word level,</td></tr><tr><td>syllable before these words as basic tone #</td><td>extended from</td></tr><tr><td/><td>training data</td></tr><tr><td>16.2 V s\u012b [\u662f, \"is\"] V: Mark the last syllable of the verb that just</td><td>Sentence pattern</td></tr><tr><td>before s\u012b as basic tone # if this verb appears again after s\u012b</td><td>level, induced</td></tr><tr><td/><td>from training data</td></tr><tr><td>16.3 che / he / chia / hia [\u9019/\u90a3(\uf9e8), \"this/that/(t)here\"] : Mark</td><td>word level</td></tr><tr><td>these words as basic tone #</td><td/></tr><tr><td>16.4 \u016b-s\u00ee [\u6709\u6642, \"sometimes\"] / put-s\u00ee [\uf967\u6642, \"from time to time\"] /</td><td>word level,</td></tr><tr><td/><td>extended from</td></tr><tr><td/><td>training data</td></tr><tr><td>] /</td><td/></tr></table>", | |
| "html": null, | |
| "num": null, | |
| "type_str": "table" | |
| }, | |
| "TABREF9": { | |
| "text": "", | |
| "content": "<table><tr><td/><td/><td colspan=\"2\">training data</td><td/><td/><td/><td colspan=\"2\">test data</td><td/></tr><tr><td>para.</td><td>no. of</td><td>no. of</td><td>no. of</td><td>accuracy</td><td>para.</td><td>no. of</td><td>no. of</td><td>no. of</td><td>accuracy</td></tr><tr><td>id.</td><td>words</td><td>syllables</td><td>errors</td><td>rate</td><td>id.</td><td>words</td><td>syllables</td><td>errors</td><td>rate</td></tr><tr><td>1</td><td>27</td><td>30</td><td>1</td><td>96.67%</td><td>1</td><td>130</td><td>184</td><td colspan=\"2\">16 91.30%</td></tr><tr><td>2</td><td>42</td><td>54</td><td>0</td><td>100.00%</td><td>2</td><td>56</td><td>85</td><td colspan=\"2\">12 85.88%</td></tr><tr><td>3</td><td>44</td><td>70</td><td>0</td><td>100.00%</td><td>3</td><td>53</td><td>84</td><td colspan=\"2\">13 84.52%</td></tr><tr><td>4</td><td>33</td><td>52</td><td>0</td><td>100.00%</td><td>4</td><td>96</td><td>143</td><td colspan=\"2\">16 88.81%</td></tr><tr><td>5</td><td>38</td><td>51</td><td>4</td><td>92.16%</td><td>5</td><td>66</td><td>97</td><td colspan=\"2\">10 89.69%</td></tr><tr><td>6</td><td>85</td><td>110</td><td>4</td><td>96.36%</td><td>6</td><td>63</td><td>86</td><td colspan=\"2\">9 89.53%</td></tr><tr><td>7</td><td>97</td><td>144</td><td>6</td><td>95.83%</td><td>7</td><td>32</td><td>43</td><td colspan=\"2\">3 93.02%</td></tr><tr><td>8</td><td>72</td><td>103</td><td>1</td><td>99.03%</td><td>8</td><td>38</td><td>58</td><td colspan=\"2\">2 96.55%</td></tr><tr><td/><td/><td/><td/><td/><td>9</td><td>122</td><td>182</td><td colspan=\"2\">25 86.26%</td></tr><tr><td colspan=\"2\">total 438</td><td>614</td><td>16</td><td colspan=\"3\">97.39% total 656</td><td>962</td><td colspan=\"2\">106 88.98%</td></tr></table>", | |
| "html": null, | |
| "num": null, | |
| "type_str": "table" | |
| }, | |
| "TABREF10": { | |
| "text": "Every syllable is affected by at least one rule, and is affected by four rules at most.", | |
| "content": "<table><tr><td/><td/><td>training Data</td><td/><td/><td>test data</td><td/></tr><tr><td>rule id.</td><td>affected syllables</td><td>accurately affected</td><td>accuracy rate</td><td>affected syllables</td><td>accurately affected</td><td>accuracy rate</td></tr><tr><td>1</td><td>614</td><td>411</td><td>66.94%</td><td>962</td><td/><td>68.81%</td></tr><tr><td>2</td><td>74</td><td>68</td><td>91.89%</td><td>112</td><td/><td>93.75%</td></tr><tr><td>3</td><td>32</td><td>24</td><td>75.00%</td><td>38</td><td/><td>68.42%</td></tr><tr><td>4</td><td>3</td><td>3</td><td>100.00%</td><td>13</td><td/><td>53.85%</td></tr><tr><td>5</td><td>65</td><td>57</td><td>87.69%</td><td>129</td><td/><td>69.77%</td></tr><tr><td>6</td><td>4</td><td>3</td><td>75.00%</td><td>4</td><td/><td>75.00%</td></tr><tr><td>7</td><td>0</td><td>0</td><td>--</td><td>0</td><td/><td>--</td></tr><tr><td>8</td><td>5</td><td>5</td><td>100.00%</td><td>3</td><td/><td>100.00%</td></tr><tr><td>9</td><td>29</td><td>29</td><td>100.00%</td><td>25</td><td/><td>100.00%</td></tr><tr><td>10</td><td>5</td><td>5</td><td>100.00%</td><td>0</td><td/><td>--</td></tr><tr><td>11</td><td>3</td><td>3</td><td>100.00%</td><td>8</td><td/><td>100.00%</td></tr><tr><td>12</td><td>8</td><td>8</td><td>100.00%</td><td>11</td><td/><td>100.00%</td></tr><tr><td>13</td><td>2</td><td>2</td><td>100.00%</td><td>6</td><td/><td>83.33%</td></tr><tr><td>14</td><td>0</td><td>0</td><td>--</td><td>0</td><td/><td>--</td></tr><tr><td>15</td><td>8</td><td>8</td><td>100.00%</td><td>6</td><td/><td>83.33%</td></tr><tr><td>16</td><td>13</td><td>13</td><td>100.00%</td><td>4</td><td/><td>100.00%</td></tr><tr><td>17</td><td>3</td><td>3</td><td>100.00%</td><td>2</td><td/><td>100.00%</td></tr><tr><td>18</td><td>9</td><td>9</td><td>100.00%</td><td>3</td><td/><td>100.00%</td></tr><tr><td>19</td><td>0</td><td>0</td><td>--</td><td>6</td><td/><td>100.00%</td></tr><tr><td>20</td><td>2</td><td>2</td><td>100.00%</td><td>0</td><td/><td>--</td></tr></table>", | |
| "html": null, | |
| "num": null, | |
| "type_str": "table" | |
| }, | |
| "TABREF11": { | |
| "text": "", | |
| "content": "<table><tr><td/><td/><td>training data</td><td/><td/><td>test data</td><td/></tr><tr><td>rule id.</td><td>no. of</td><td>no. of</td><td>accuracy</td><td>no. of</td><td>no. of</td><td>accuracy</td></tr><tr><td/><td>dominant</td><td>accurate</td><td>rate</td><td>dominant</td><td>accurate</td><td>rate</td></tr><tr><td/><td>rule</td><td>dominant</td><td/><td>rule</td><td>dominant</td><td/></tr><tr><td/><td/><td>rule</td><td/><td/><td>rule</td><td/></tr><tr><td>1</td><td>381</td><td>371</td><td>97.38%</td><td>616</td><td/><td>92.21%</td></tr><tr><td>2</td><td>62</td><td>62</td><td>100.00%</td><td>104</td><td/><td>95.19%</td></tr></table>", | |
| "html": null, | |
| "num": null, | |
| "type_str": "table" | |
| }, | |
| "TABREF12": { | |
| "text": "", | |
| "content": "<table><tr><td>Errors</td><td>Possible Solutions</td></tr><tr><td>(a) Due to dictionary limitation (not having the</td><td>Increase entries</td></tr><tr><td>words)</td><td/></tr><tr><td>(b) Due to lack of punctuation marks</td><td>Pre-process, but this is very difficult</td></tr><tr><td colspan=\"2\">(c) Due to wrong POS because of homonymy Apply semantic knowledge</td></tr><tr><td>(d) Due to indeterminate POS or multiple</td><td>Tagging disambiguity</td></tr><tr><td>candidates</td><td/></tr><tr><td>(e) Caused by inconsistent orthography in</td><td>Pre-process the sources or deal with the</td></tr><tr><td>hyphen segmentation</td><td>procedures of adding or removing hyphens</td></tr><tr><td/><td>automatically</td></tr><tr><td>(f) Due to incomplete sandhi rule set</td><td>Refine the sandhi rules while avoiding side</td></tr><tr><td/><td>effects</td></tr><tr><td>(g) Associated with quantitative words;</td><td>Add DM rules</td></tr><tr><td>(h) Associated with proper nouns</td><td>Detect proper nouns</td></tr><tr><td>(i) Associated with sentence pattern</td><td>Add sandhi rules for sentence patterns</td></tr><tr><td>(j) Possibly other sources of error yet to be</td><td/></tr><tr><td>identified</td><td/></tr></table>", | |
| "html": null, | |
| "num": null, | |
| "type_str": "table" | |
| } | |
| } | |
| } | |
| } |