| { |
| "paper_id": "Y16-2008", |
| "header": { |
| "generated_with": "S2ORC 1.0.0", |
| "date_generated": "2023-01-19T13:47:28.195027Z" |
| }, |
| "title": "Developing an Unsupervised Grammar Checker for Filipino Using Hybrid N-grams as Grammar Rules", |
| "authors": [ |
| { |
| "first": "Matthew", |
| "middle": [ |
| "Phillip" |
| ], |
| "last": "Go", |
| "suffix": "", |
| "affiliation": {}, |
| "email": "" |
| }, |
| { |
| "first": "Allan", |
| "middle": [], |
| "last": "Borra", |
| "suffix": "", |
| "affiliation": {}, |
| "email": "allan.borra@dlsu.edu.ph" |
| } |
| ], |
| "year": "", |
| "venue": null, |
| "identifiers": {}, |
| "abstract": "This study focuses on using hybrid n-grams as grammar rules for detecting grammatical errors and providing corrections in Filipino. These grammar rules are derived from grammatically-correct and tagged texts which are made up of part-of-speech (POS) tags, lemmas, and surface words sequences. Due to the structure of the rules used by this system, it presents an opportunity to have an unsupervised grammar checker for Filipino when coupled with existing POS taggers and morphological analyzers. The approach is also customized to cover different error types present in the Filipino language. The system achieved 82% accuracy when tested on checking erroneous and error-free texts.", |
| "pdf_parse": { |
| "paper_id": "Y16-2008", |
| "_pdf_hash": "", |
| "abstract": [ |
| { |
| "text": "This study focuses on using hybrid n-grams as grammar rules for detecting grammatical errors and providing corrections in Filipino. These grammar rules are derived from grammatically-correct and tagged texts which are made up of part-of-speech (POS) tags, lemmas, and surface words sequences. Due to the structure of the rules used by this system, it presents an opportunity to have an unsupervised grammar checker for Filipino when coupled with existing POS taggers and morphological analyzers. The approach is also customized to cover different error types present in the Filipino language. The system achieved 82% accuracy when tested on checking erroneous and error-free texts.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Abstract", |
| "sec_num": null |
| } |
| ], |
| "body_text": [ |
| { |
| "text": "According to the philosopher and educator Kevin Browne, poor grammar implies two negative sentiments towards the writer: either he is not intelligent or he just does not care about his writing any better. Backing on this problem, there has been many researches and advances in the field of computer-aided grammar checking such as Microsoft Word, Google Docs, Grammarly, LanguageTool, and Ginger. These software solutions can detect syntactical errors such as spelling, punctuation, word forms, and word usages. However, most of these solutions have focused on the English language. There has been very few works in the Filipino language despite being a language of at least 100 million people 1 . Additionally, it is difficult to use an existing grammar checker system of one language and apply it on another since the system would have its specific design and 1 http://www.philstar.com/headlines/2016/01/03/1538653/ph ilippines-population-seen-hit-104m functionalities tackling the unique phenomena of its target language.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Introduction", |
| "sec_num": "1." |
| }, |
| { |
| "text": "The Filipino language, just like any other language, has its own unique phenomena which serve as a challenge in developing its own grammar checker system. It has a 'large vocabulary of root, borrowed, and derived words' caused by the arrival and/or colonization of foreign countries including: Spain, USA, and China in the Filipino land 2 . It also has a high degree of inflection and uses variety of affixes to change the part-of-speech of a root word (ex. root: tira 'live [on a house]', tira + han = tirahan 'house') or change the focus and aspect of a verb (tirhan 'live' -neutral aspect/object focus, titira 'will live' -contemplative aspect/ actor focus, tumira 'lived' -perfective aspect/ actor focus. Another linguistic phenomenon in Filipino is its free-word order structure. Filipino sentences, in its natural form, follow the predicate-subject sentence format (ex. Masaya ako -word-perword is translated as 'Happy I') or as subjectpredicate sentence format (ex. Ako ay masayaword-per-word is translated as 'I [none] happy') where the word ay acts as a lexical marker and is usually placed after the subject and before the predicate. In the Filipino language, direct objects, adjectives and adverbs may also be written as phrases and including prepositional phrases, they also follow the free-word order and not being limited to just one position in the sentence (Ramos, 1971) . For example, the sentence 'Mark ate an apple.' can be translated to: Si Mark ay kumain ng mansanas., Kumain si Mark ng mansanas., and Kumain ng mansanas si Mark. As seen in the last two translations, the direct object phrase ng mansanas 'apple' can be placed directly after the verb or after the subject yet both produce the exact same meaning.", |
| "cite_spans": [ |
| { |
| "start": 1373, |
| "end": 1386, |
| "text": "(Ramos, 1971)", |
| "ref_id": "BIBREF11" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Introduction", |
| "sec_num": "1." |
| }, |
| { |
| "text": "As of this writing, there are still no grammar-checking software systems for Filipino that is publicly available that cover broad-range of grammatical errors. This fact may be associated with the complex structure of the Filipino language which makes it difficult in constructing (error) grammar rules. Among the few existing grammar checkers in Filipino are: Panuring Pampanitikan (PanPam) by Jasa et al. (2007) and Language Tool for Filipino (LTF) by Oco & Borra (2011) . PanPam is a syntax and semantics-based grammar checker for Filipino that makes use of error patterns as rules and lexical functional grammar as its parsing algorithm. LTF, on the other hand, uses a rule file containing error patterns in the form of regular expressions and part-of-speech tags and a dictionary file in detecting its errors and providing corresponding suggestions. Although these systems, especially LTF, could distinctly recognize grammatical errors from correct text by using error patterns, the main concern with these systems is that the parser rules, dictionaries, affix-to-root-word mappings, wordto-part-of-speech mappings, error patterns, and other files are manually defined which is a very tedious task to cover the entire language and all possible errors in it especially that the language is ever growing and the number of errors committed by writers are directly proportional to it. This concern is evident on the systems' presented limitations and results where only a small subset of errors was covered.", |
| "cite_spans": [ |
| { |
| "start": 394, |
| "end": 412, |
| "text": "Jasa et al. (2007)", |
| "ref_id": null |
| }, |
| { |
| "start": 453, |
| "end": 471, |
| "text": "Oco & Borra (2011)", |
| "ref_id": "BIBREF6" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Introduction", |
| "sec_num": "1." |
| }, |
| { |
| "text": "In other languages such as English, there are existing works such as Lexbar (Tsao & Wible, 2009) , EdIt (Huang et al., 2011) , Google books n-gram corpus as grammar checker (Nazar & Renau, 2012) , and Chunk-based grammar checker for translated sentences (Lin et al., 2011) which are unsupervised grammar checker systems that make use of grammatically correct texts, their corresponding part-of-speech (POS) tags, and/or lemmas converted into n-gram sequences and used as grammar rules.", |
| "cite_spans": [ |
| { |
| "start": 76, |
| "end": 96, |
| "text": "(Tsao & Wible, 2009)", |
| "ref_id": "BIBREF7" |
| }, |
| { |
| "start": 104, |
| "end": 124, |
| "text": "(Huang et al., 2011)", |
| "ref_id": "BIBREF0" |
| }, |
| { |
| "start": 173, |
| "end": 194, |
| "text": "(Nazar & Renau, 2012)", |
| "ref_id": "BIBREF10" |
| }, |
| { |
| "start": 254, |
| "end": 272, |
| "text": "(Lin et al., 2011)", |
| "ref_id": "BIBREF8" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Introduction", |
| "sec_num": "1." |
| }, |
| { |
| "text": "The Lexbar application (Tsao & Wible, 2009 ) generated hybrid n-grams, which are ngrams composed of words, POS tags, and lemmas. These hybrid n-grams are generated from actual tagged word sequences. For example, given phrases such as 'from her point of view' and 'from his point of view', the system will be able to generate the hybrid rule 'from [dps] 3 point of view'. This rule can be used to flag the phrase 'from my point of view' as grammatically correct and the phrase 'from him point of view' as incorrect. The Lexbar app was only tested on substitution-correctable errors. The EdIt system (Huang et al., 2011) This research aims to build an unsupervised grammar checker system for Filipino using hybrid n-grams as grammar rules following a similar format as Lexbar's grammar rules. These rules will be used to detect grammatical errors in Filipino and provide suggestions such as substitution, insertion, deletion, merging, and unmerging extending the existing suggestions made by both Lexbar and EdIt.", |
| "cite_spans": [ |
| { |
| "start": 23, |
| "end": 42, |
| "text": "(Tsao & Wible, 2009", |
| "ref_id": "BIBREF7" |
| }, |
| { |
| "start": 598, |
| "end": 618, |
| "text": "(Huang et al., 2011)", |
| "ref_id": "BIBREF0" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Introduction", |
| "sec_num": "1." |
| }, |
| { |
| "text": "Aside from the free-word order structure in Filipino, there are other linguistic phenomena such as being morphologically rich, existence of compound words, and the rule in Filipino: \"Kung ano ang bigkas, siyang sulat\" 'Spell as you pronounce it' (Ortograpiyang Pambansa, 2013) .", |
| "cite_spans": [ |
| { |
| "start": 261, |
| "end": 276, |
| "text": "Pambansa, 2013)", |
| "ref_id": "BIBREF9" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Filipino Linguistic Phenomena", |
| "sec_num": "2." |
| }, |
| { |
| "text": "There are at least 50 affixes and other morphologies such as partial reduplication, full reduplication, and compounding that are used in Filipino. These morphologies are categorized into three: inflectional -changes in word form that 'accompany case, gender, number, tense, person, mood, or voice that have no effect in the word's part-of-speech'; derivational -changes in word form that changes the word's part-ofspeech category; and compounding -'where independent words are concatenated in some way to form a new word' (Bonus, 2003) . See Table 1 for some of the different forms of the root word kain 'eat'. There are also affixes that are separated by a hyphen (-) from its root word or morpheme (ex. mang-akit 'to entice' from the root akit 'entice'). There are also cases wherein addition or insertion of an affix to a word could alter the spelling of its base form (ex. The prefix pang-+ palit 'change' = pamalit 'item for changing'). However, not all affixes and reduplication can be applied to any word. For instance, the root word luto 'cook' can use 'nag-' as prefix but kain 'eat' cannot. It should also be noted that there are assimilated words from English in Filipino wherein affixes are also appended to it (ex. magce-cellphone 'will use a cellphone', i-file 'to file (a document)'). The Filipino language also has its own set of compound words. There are two ways to combine words together, either with the use of a hyphen (ex. halo-halo '(a type of Filipino dessert)' from the word halo 'mix', and kisapmata 'instant' from the words kisap 'blink' & mata 'eye') or just combining them as is (ex. kapitbahay 'neighbor' from the words kapit 'hold onto' & bahay 'house', and hanapbuhay 'livelihood' from the words hanap 'find' & buhay 'life') (Paz, 2003) .", |
| "cite_spans": [ |
| { |
| "start": 522, |
| "end": 535, |
| "text": "(Bonus, 2003)", |
| "ref_id": "BIBREF3" |
| }, |
| { |
| "start": 1757, |
| "end": 1768, |
| "text": "(Paz, 2003)", |
| "ref_id": "BIBREF1" |
| } |
| ], |
| "ref_spans": [ |
| { |
| "start": 542, |
| "end": 549, |
| "text": "Table 1", |
| "ref_id": null |
| } |
| ], |
| "eq_spans": [], |
| "section": "Filipino Linguistic Phenomena", |
| "sec_num": "2." |
| }, |
| { |
| "text": "Another important linguistic phenomenon in Filipino is the rule: \"Kung ano ang bigkas, siyang sulat\" 'Spell as you pronounce it' (Ortograpiyang Pambansa, 2013) . As the rule states, the words in Filipino are usually spelled as they are pronounced with some exceptions. This phenomenon simplifies the way Filipino words are spelled out (ex. Filipinized form of 'computer' as kompyuter) but also causes some spelling confusion which will be discussed in the next section.", |
| "cite_spans": [ |
| { |
| "start": 144, |
| "end": 159, |
| "text": "Pambansa, 2013)", |
| "ref_id": "BIBREF9" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Word", |
| "sec_num": null |
| }, |
| { |
| "text": "In understanding the error types that exist in Filipino writing, three references were used: The Cambridge Learner Corpus (Nicholls, 1999) , Wikapedia (2015) , and a parallel corpus of 1252 erroneous-and-correct word and phrase pairs from sentences written by Filipino university students.", |
| "cite_spans": [ |
| { |
| "start": 122, |
| "end": 138, |
| "text": "(Nicholls, 1999)", |
| "ref_id": "BIBREF2" |
| }, |
| { |
| "start": 141, |
| "end": 157, |
| "text": "Wikapedia (2015)", |
| "ref_id": "BIBREF14" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Error Types", |
| "sec_num": "3." |
| }, |
| { |
| "text": "The Cambridge Learner Corpus contains 16 million words from English examination scripts by learners of English containing different types of errors. The corpus categorized the error types into general and specific errors. The proponents noticed that some error categories would have its Filipino counterpart such as wrong form used, missing word/phrase, word/phrase needs replacing, unnecessary word/phrase, punctuation errors, countability errors, determiner agreement, incorrect verb inflection, spelling errors, and other error categories also exist in Filipino.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Error Types", |
| "sec_num": "3." |
| }, |
| { |
| "text": "Wikapedia (2015) is a booklet created by the Presidential Communications Development and Strategic Planning Office of the Philippines containing correct usage of affixes, words, and phrases in Filipino which people may find confusing. One example described in the book would be the use of ng, a function word defining possession (ex. aso ng kapitbahay 'dog of neighbor') and in a direct object phrase (ex. kumain ng mansanas 'ate an apple') vs the use of nang which is commonly used before an adverb (ex. kumain nang mabilis 'ate fast'). The usage of these two words is confusing because it is pronounced almost exactly the same. Other examples contained in the booklet are proper usage of affixes and words, morphophonemics, usage of hyphens and spaces, and others.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Error Types", |
| "sec_num": "3." |
| }, |
| { |
| "text": "After analyzing the parallel corpus of 1252 erroneous-correct word/phrase pairs, it is found that majority of the errors fall under spelling errors, incorrect usage of affixes/reduplication which is mostly caused by usage of hyphens and spaces, and wrong word usage.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Error Types", |
| "sec_num": "3." |
| }, |
| { |
| "text": "It is observed that one reason the students made spelling errors is because of the way a word is pronounced which is usually simplified for conversational use. Some of these simplified words, see Table 2 , are still not accepted in formal Filipino writing which cause spelling errors. Another cause of spelling errors is the confusion whether to spell an English borrowed word in its English version or convert it to its Filipinized spelling version.", |
| "cite_spans": [], |
| "ref_spans": [ |
| { |
| "start": 196, |
| "end": 203, |
| "text": "Table 2", |
| "ref_id": "TABREF3" |
| } |
| ], |
| "eq_spans": [], |
| "section": "Error Types", |
| "sec_num": "3." |
| }, |
| { |
| "text": "There were many instances of affix errors where the students were confused whether a word is an affix of a word, a separate word, or if there should be a hyphen between the affix and the root word. A few of the affix errors also show the confusion of students in selecting an appropriate affix of a verb when used for a certain focus and/or aspect. See Table 3 .", |
| "cite_spans": [], |
| "ref_spans": [ |
| { |
| "start": 353, |
| "end": 360, |
| "text": "Table 3", |
| "ref_id": null |
| } |
| ], |
| "eq_spans": [], |
| "section": "Error Types", |
| "sec_num": "3." |
| }, |
| { |
| "text": "The students also committed several mistakes in identifying which word to use in certain situations which is caused by unfamiliarity with Filipino syntax rules. See Table 4 .", |
| "cite_spans": [], |
| "ref_spans": [ |
| { |
| "start": 165, |
| "end": 172, |
| "text": "Table 4", |
| "ref_id": "TABREF4" |
| } |
| ], |
| "eq_spans": [], |
| "section": "Error Types", |
| "sec_num": "3." |
| }, |
| { |
| "text": "Other errors that exist in the parallel corpus include the lack of space between words (ex. pa rin 'still' incorrectly written as parin), compound words that was separated by a space (ex. arawaraw 'everyday' incorrectly written as araw araw) and punctuation errors where some commas or periods are missing. ", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Error Types", |
| "sec_num": "3." |
| }, |
| { |
| "text": "The grammar checker named Gramatika that is discussed in this paper utilizes the existing implementation of the Lexbar application by Tsao & Wible (2009) and extends it to cover more error types, some of which are unique in the Filipino language. It uses n-grams as rules, commonly referred to as hybrid n-grams, from grammatically correct texts consisting of words, POS tags, and lemmas to detect grammatical errors and provide suggestions containing possible corrections. The production of POS tags, and lemmas can be produced by existing POS taggers and morphological analyzers 6 for Filipino making the system unsupervised such that new grammatically correct texts can be fed through these systems and to Gramatika to easily increase the number of grammar rules.", |
| "cite_spans": [ |
| { |
| "start": 134, |
| "end": 153, |
| "text": "Tsao & Wible (2009)", |
| "ref_id": "BIBREF7" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Overview of the Grammar Checker", |
| "sec_num": "4." |
| }, |
| { |
| "text": "Even though Gramatika also uses hybrid n-grams similar to Lexbar's (Tsao & Wible, 2009) and slightly similar to EdIt's (Huang et al., 2011) , the approach in deriving the hybrid n-grams is different. Gramatika uses a clustering approach as opposed to Lexbar's pruning and EdIt's collocations-based approaches. The n-gram sizes used as rules range from 2 to 7. For example, given an incorrect phrase para sa bata ang laruan ni iyon. '?that? toy is for the kid', if Gramatika has the hybrid 7-gram 'para_sa [NNC] [DTC] [NNC] na [PRO] .' 7 , then it can immediately suggest to change the word ni '(a grammatical particle used before a personal proper noun)' to na '(a grammatical particle used around adjectives, pointing pronouns, and others)' which produces the corrected version: para sa bata ang laruan na iyon 'that toy is for the kid' which is a more appropriate suggestion than the suggestion produced by the trigram [NNC] ni [NNP] 8 to change iyon to a proper noun (ex. Mark) producing the corrected version: para sa bata ang laruan ni Mark 'Mark's toy is for the kid'. The use of larger ngram sizes increases the context from which a suggestion can be based from.", |
| "cite_spans": [ |
| { |
| "start": 67, |
| "end": 87, |
| "text": "(Tsao & Wible, 2009)", |
| "ref_id": "BIBREF7" |
| }, |
| { |
| "start": 119, |
| "end": 139, |
| "text": "(Huang et al., 2011)", |
| "ref_id": "BIBREF0" |
| }, |
| { |
| "start": 511, |
| "end": 516, |
| "text": "[DTC]", |
| "ref_id": null |
| }, |
| { |
| "start": 526, |
| "end": 531, |
| "text": "[PRO]", |
| "ref_id": null |
| }, |
| { |
| "start": 921, |
| "end": 926, |
| "text": "[NNC]", |
| "ref_id": null |
| }, |
| { |
| "start": 930, |
| "end": 935, |
| "text": "[NNP]", |
| "ref_id": null |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Rules Learning", |
| "sec_num": "4.1" |
| }, |
| { |
| "text": "In the clustering approach, all n-gram sequences are retrieved from grammatically correct texts and are stored in the database. During the storing process, the frequency of all POS tag sequences is counted. POS tag sequences exceeding the threshold of 2 are retrieved and the word n-grams are grouped as clusters. For each n-gram clusters, the module checks if there are any token slot that can be generalized to POS level. For example, if a cluster has the instances nagpunta sa bayan 'went to the town' and bumisita sa bahay 'visited the house', the first and third tokens can be generalized because it meets the minimum difference threshold of 2. This produces the hybrid n-gram [VBTS] sa [NNC] which can be used to flag the phrase umupo sa silya 'sat on the chair' as grammatically correct or used to detect grammatical errors. The n-gram rules are stored in the database as sequences of words, POS tags, lemmas, and a Boolean sequence denoting which token slots are generalized. This is done to allow Gramatika to provide word-specific suggestions 7 Based from the Rabo & Cheng (2006) tag set, NNC = common noun, DTC = determiner for common nouns, PRO = pronoun pointing to an object 8 NNP = proper noun and to also identify the appropriate transformed word to a specific POS -lemma mapping.", |
| "cite_spans": [ |
| { |
| "start": 692, |
| "end": 697, |
| "text": "[NNC]", |
| "ref_id": null |
| }, |
| { |
| "start": 1053, |
| "end": 1054, |
| "text": "7", |
| "ref_id": null |
| }, |
| { |
| "start": 1070, |
| "end": 1089, |
| "text": "Rabo & Cheng (2006)", |
| "ref_id": null |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Rules Learning", |
| "sec_num": "4.1" |
| }, |
| { |
| "text": "In detecting grammatical errors and producing suggestions based on the hybrid n-grams, a weighted Levenshtein edit distance algorithm is used. This algorithm is commonly used in spell checking to compute how many edits it will take to convert a potentially misspelled word to a correct word in the dictionary. It has also been used by EdIt (Huang et al., 2011) in providing corrections by substitution, insertion, and deletion. In Gramatika, the edit distance algorithm is extended to detect errors and provide suggestions correctable by substitution, insertion, deletion, spelling correction, unmerging, and merging. The error types that exists in Filipino are grouped based on the six suggestion types, see Table 5 . In producing suggestions, Gramatika parses the input, which is POS and lemma-tagged, into n-grams starting from size 7 down to 2. For each input n-gram, it retrieves hybrid n-gram rules \"similar\" to the input n-gram from the database. A rule is considered \"similar\" to an input n-gram if at least n-2 POS tokens of it are equal to the POS tokens in the input n-gram. Three sizes of the rules are also retrieved for each input n-gram: rules that are of equal size to the input n-gram to be used for substitution and spelling correction suggestions, rules that are one token size larger to produce insertion and unmerging suggestions, and rules that are one token size smaller to produce deletion and merging suggestions. If an input n-gram is \"equal\" to a rule n-gram of the same size, then it is considered grammatically correct, which is denoted by an edit distance value of 0. \"Equal\" tokens, in this context, are defined as tokens having the same POS tag if the POS tag of the rule n-gram token is generalized or tokens that are equal in surface word level.", |
| "cite_spans": [ |
| { |
| "start": 340, |
| "end": 360, |
| "text": "(Huang et al., 2011)", |
| "ref_id": "BIBREF0" |
| } |
| ], |
| "ref_spans": [ |
| { |
| "start": 709, |
| "end": 716, |
| "text": "Table 5", |
| "ref_id": null |
| } |
| ], |
| "eq_spans": [], |
| "section": "Error Detection", |
| "sec_num": "4.2" |
| }, |
| { |
| "text": "A substitution suggestion is outputted when all tokens in the rule n-gram except one are equal to its respective tokens in the input n-gram in the same index. The unequal token is categorized depending on its error type: affix errors or wrong word usage. In detecting affix errors, the system checks if the lemma of the unequal input token is equal to the respective rule token. If the lemmas are equal but the words and/or POS tags are not equal, then it is assumed that there is an affix usage error with respect to the rule n-gram used. For example, given the input kumain siya bukas 'he ate tomorrow' is compared against the rule ngram [VBTF] [PRS] bukas 9 , where one instance of the POS tag [VBTF] is the word kakain 'will eat', since kumain 'ate' and kakain has the same lemma kain 'eat', the error is classified as an affix usage error which will produce a suggestion 'Change kumain to kakain' to produce the phrase kakain siya bukas 'He will eat tomorrow'. For wrong word usage errors, Gramatika suggests that the unequal rule token should replace the unequal input token. Consider the example: Bumili si bata ng laruan '?the? child bought a toy'. It is considered grammatically incorrect because the word si is used as a determiner for proper nouns. Using the trigram rule [VBTS] ang [NNC] 10 , the correct determiner for common nouns ang will be suggested to replace the incorrect word si. Corresponding edit distance values will be assigned to the outputted substitution suggestions as seen in Table 6 A spelling suggestion is similar to the substitution suggestion criteria where in all tokens except one should be equal to its respective tokens. The unequal tokens are compared using a character-level edit distance algorithm. If it meets the spelling difference threshold, then the token in the rule n-gram is outputted as a spelling correction suggestion. This approach is slightly similar to the traditional way of spell checking which is dictionary look up, but this uses context in providing an appropriate suggestion. For example, given the input Bakt ?, using dictionary look-up alone, two possible suggestions can be produced: Bakat 'a mark/trace' or Bakit 'why'. Using a contextbased approach and the rule [PRQ] ? 11 where the word bakit is an instance of the POS tag [PRQ], Gramatika will suggest the word Bakit to form the correct sentence Bakit? 'Why?'. This suggestion is assigned an edit distance value of 0.65, as seen in Table 6 .", |
| "cite_spans": [], |
| "ref_spans": [ |
| { |
| "start": 1506, |
| "end": 1513, |
| "text": "Table 6", |
| "ref_id": "TABREF7" |
| }, |
| { |
| "start": 2451, |
| "end": 2458, |
| "text": "Table 6", |
| "ref_id": "TABREF7" |
| } |
| ], |
| "eq_spans": [], |
| "section": "Correction Error Types", |
| "sec_num": null |
| }, |
| { |
| "text": "An insertion suggestion is outputted when all tokens in the input n-gram find their corresponding equal tokens in the rule n-gram and one token from the rule n-gram does not have a matched token. For example, given the input n-gram a b d and rule n-gram a b c d, tokens a, b, and d are equal tokens and c does not have a match which is outputted as an insertion suggestion. This suggestion is given an edit distance value of 1.0, as seen in Table 6. A deletion suggestion is outputted when all tokens in the rule n-gram find their corresponding equal tokens in the input n-gram and one token from the input n-gram does not have a matched token. This one token will be suggested to be deleted. This suggestion is also given an edit distance value of 1.0, as seen in Table 6 .", |
| "cite_spans": [], |
| "ref_spans": [ |
| { |
| "start": 441, |
| "end": 449, |
| "text": "Table 6.", |
| "ref_id": "TABREF7" |
| }, |
| { |
| "start": 765, |
| "end": 772, |
| "text": "Table 6", |
| "ref_id": "TABREF7" |
| } |
| ], |
| "eq_spans": [], |
| "section": "Correction Error Types", |
| "sec_num": null |
| }, |
| { |
| "text": "An unmerging suggestion is outputted when n-1 tokens in the input n-gram are equal to its respective tokens in the rule n-gram leaving one token in the input n-gram and two adjacent tokens in the rule n-gram without matching equal tokens. The system concatenates the two adjacent tokens in the rule n-gram together using a hyphen or removal of the space and checks if it equates to the spelling of the input token. If equal, then an unmerging suggestion is produced. For example, given the input four-gram a b cd e, and rule five-gram a b c d e, the tokens a, b, and e are equal tokens. After which, the tokens c and d in the rule n-gram is concatenated, since it equals the token cd in the input n-gram, then the unmerging suggestion is outputted. This suggestion is given an edit distance value of 0.7, as seen in Table 6. A merging suggestion is outputted when n -1 tokens in the rule n-gram are equal to its respective tokens in the input n-gram leaving one token in the rule n-gram and two adjacent tokens in the input n-gram without matching equal tokens. The system concatenates the two adjacent tokens in the input n-gram together using a hyphen or removal of the space and checks if it equates to the spelling of the rule token. If yes, then a merging suggestion is produced. For example, given the input five-gram a b c d e, and rule four-gram a b cd e, the tokens a, b, and e are equal tokens. After which, the tokens c and d in the input n-gram is concatenated, since it equals the token cd in the rule n-gram, then a merging suggestion is outputted. This suggestion is also given an edit distance value of 0.7, as seen in Table 6 .", |
| "cite_spans": [], |
| "ref_spans": [ |
| { |
| "start": 816, |
| "end": 824, |
| "text": "Table 6.", |
| "ref_id": "TABREF7" |
| }, |
| { |
| "start": 1635, |
| "end": 1642, |
| "text": "Table 6", |
| "ref_id": "TABREF7" |
| } |
| ], |
| "eq_spans": [], |
| "section": "Correction Error Types", |
| "sec_num": null |
| }, |
| { |
| "text": "It should be noted that the edit distance values used are arbitrary and is mainly used for prioritization of suggestions only as the edit distance threshold is set to 1. In cases that there would be ties in terms of edit distance values (ex. three wrong word different POS tag suggestions), the frequency of how many times each suggestion is produced by the n-gram rules is also considered.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Correction Error Types", |
| "sec_num": null |
| }, |
| { |
| "text": "The Gramatika system is tested on 70 phrases (35 erroneous and 35 error-free) retrieved from PanPam (Jasa et al., 2007 ) Wikapedia (2015 , translated documents by Filipino university students, and news articles. A small corpus of 2,668 complex sentences which consists of 70,312 tokens (14,575 unique tokens) is used in training of the n-gram rules which result to 83% accuracy, 93% precision, and 71% recall. Table 7 shows a summary of figures. 18 out of 25 erroneous phrases were marked as erroneous, and 23 out of 25 error-free phrases were marked as error-free. This result shows significant potential for an unsupervised grammar checker that currently only uses a small corpus of grammatically correct sentences.", |
| "cite_spans": [ |
| { |
| "start": 100, |
| "end": 118, |
| "text": "(Jasa et al., 2007", |
| "ref_id": null |
| }, |
| { |
| "start": 119, |
| "end": 136, |
| "text": ") Wikapedia (2015", |
| "ref_id": "BIBREF14" |
| } |
| ], |
| "ref_spans": [ |
| { |
| "start": 410, |
| "end": 417, |
| "text": "Table 7", |
| "ref_id": null |
| } |
| ], |
| "eq_spans": [], |
| "section": "Results and Analysis", |
| "sec_num": "5." |
| }, |
| { |
| "text": "In detecting errors, the system was able to produce word-specific suggestions for most errors except for one instance: Noong 2006, mananalo ang 'Last 2006, (something/someone) will win', having a verb aspect/tense and adverb of time disagreement, where the system only suggest to replace the word mananalo 'will win' with any perfective verb [VBTS] because it did not have the word-POS-lemma mapping nanalo 'won'- [VBTS] panalo 'win'. Other detected errors with produced word suggestions include affix errors (ex. linaan -root word laan 'allot' cannot use the infix '-in-' -> naglaan 'alloted'), wrong word usages (ex. nang vs ng, para_sa vs para_kay), spelling errors (ex. kikumpara -> kinukumpara 'comparing', lalake -> lalaki 'man', nag-simula -> nagsimula 'started', nagka-loob -> nagkaloob 'given', skolar -> iskolar 'scholar', pwesto -> puwesto 'position'), merging/unmerging errors (ex maka kuha -> makakuha 'to get', parin -> pa rin 'still'), and missing punctuation (ex. Mayo 31 2016 -> Mayo 31, 2016 'May 31, 2016 .", |
| "cite_spans": [ |
| { |
| "start": 414, |
| "end": 420, |
| "text": "[VBTS]", |
| "ref_id": null |
| }, |
| { |
| "start": 975, |
| "end": 992, |
| "text": "(ex. Mayo 31 2016", |
| "ref_id": null |
| }, |
| { |
| "start": 993, |
| "end": 1009, |
| "text": "-> Mayo 31, 2016", |
| "ref_id": null |
| }, |
| { |
| "start": 1010, |
| "end": 1023, |
| "text": "'May 31, 2016", |
| "ref_id": null |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Sentences", |
| "sec_num": null |
| }, |
| { |
| "text": "Gramatika failed to produce correct suggestions for some erroneous inputs because of several reasons. One reason is that some tags in the Rabo & Cheng (2006) tag set are too generalized. For instance, the sequence para sa Mark 'for the Mark' was incorrectly flagged as grammatically correct which is supposedly para kay Mark 'for Mark' because para sa is used for common nouns and proper nouns of place. This occurred because all proper nouns only use the tag [NNP] and an n-gram rule para_sa [NNP] was produced from phrases with proper nouns of place such as para sa Amerika, and para sa Taiwan. Another reason is that some words or phrase sequences were not part of the training corpus. For example, the phrase taga Manila siya was not corrected to taga-Manila siya 'He's from Manila' because the word taga-Manila was not in the database, thus preventing a merging suggestion. The system also failed to detect the grammatical error in the phrase Maganda si. 'is beautiful.' which can be corrected as Maganda si Maria. 'Maria is beautiful.' because the corpus sentences were all complex sentences and thus the correction for the simple sentence was not produced. The system also incorrectly flagged the phrase na bibigay ang where bibigay is an invalid contemplative form [VBTF] of the word bigay 'give' which is supposedly magbibigay 'will give' because there is no magbibigay-[VBTF]-bigay mapping in the database. This is caused by the word bibigay being given the tag [VBTF] since it follows the partial reduplication rule for the said tag and the hybrid sequence na [VBTF] ang being considered grammaticallycorrect based from sequences such as na gagampanan ang 'will take the role of'.", |
| "cite_spans": [ |
| { |
| "start": 138, |
| "end": 157, |
| "text": "Rabo & Cheng (2006)", |
| "ref_id": null |
| }, |
| { |
| "start": 460, |
| "end": 465, |
| "text": "[NNP]", |
| "ref_id": null |
| }, |
| { |
| "start": 493, |
| "end": 498, |
| "text": "[NNP]", |
| "ref_id": null |
| }, |
| { |
| "start": 1273, |
| "end": 1279, |
| "text": "[VBTF]", |
| "ref_id": null |
| }, |
| { |
| "start": 1571, |
| "end": 1577, |
| "text": "[VBTF]", |
| "ref_id": null |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Sentences", |
| "sec_num": null |
| }, |
| { |
| "text": "The system performed well in correctly flagging error-free phrases even though many of the words were not in the training corpus. This result can be attributed to the POS tags of the input phrases and the generalized rules. For instance, the phrase ang pagtakbo ng mayor 'mayor's candidacy' was flagged correctly because of the n-gram rule ang [NNC] ng [NNC] that came from instances such as ang kulay ng apoy 'fire's color', ang proseso ng produksyon 'production's process'. However, there are two instances wherein the system incorrectly flagged error-free phrases: Siya ay masaya. 'She is happy' because simple sentences that contain the word ay did not occur in the training corpus and Kinuha ko ito dito. 'I got it from here' because neither the word sequence ko ito nor the POS sequence [PRS] [PRO] is not in the n-gram rules.", |
| "cite_spans": [ |
| { |
| "start": 344, |
| "end": 349, |
| "text": "[NNC]", |
| "ref_id": null |
| }, |
| { |
| "start": 353, |
| "end": 358, |
| "text": "[NNC]", |
| "ref_id": null |
| }, |
| { |
| "start": 799, |
| "end": 804, |
| "text": "[PRO]", |
| "ref_id": null |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Sentences", |
| "sec_num": null |
| }, |
| { |
| "text": "This paper presents an unsupervised grammar checker system for Filipino that uses grammatically correct texts as grammar rules in the form of hybrid n-grams. It achieved 83% accuracy, 93% precision, and 71% recall in detecting broad-range of grammar errors in Filipino using only a small corpus of 2,668 sentences which can potentially be further improved as the size and variety of training sentences increases.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Summary & Future Works", |
| "sec_num": "6." |
| }, |
| { |
| "text": "Future works for this system includes exploring other tag sets or modifying existing ones to create more specific tag groups for the system to avoid errors caused by tag groups that are too general. Corpus size may also be increased to produce more n-gram rules and word-POS-lemma mappings used for grammar checking.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Summary & Future Works", |
| "sec_num": "6." |
| }, |
| { |
| "text": "http://ffemagazine.com/the-origin-of-the-filipinolanguage-wikang-filipino/", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "", |
| "sec_num": null |
| }, |
| { |
| "text": "30th Pacific Asia Conference on Language, Information and Computation (PACLIC 30)Seoul, Republic of Korea, October 28-30, 2016", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "", |
| "sec_num": null |
| }, |
| { |
| "text": "dps is the part-of-speech (POS) tag for possessive pronouns such as his, her, my, their, etc in the CLAWS5 tagset. 4 V-ing is the POS tag for verbs followed by -ing in the CLAWS5 tagset. 5 nn0 is the POS tag for neutral nouns in the CLAWS5 tagset.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "", |
| "sec_num": null |
| }, |
| { |
| "text": "See Rabo & Cheng (2006) andBonus (2003)", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "", |
| "sec_num": null |
| }, |
| { |
| "text": "PRQ = question/interrogative pronoun", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "", |
| "sec_num": null |
| } |
| ], |
| "back_matter": [], |
| "bib_entries": { |
| "BIBREF0": { |
| "ref_id": "b0", |
| "title": "EdIt: A Broad-coverage Grammar Checker Using Pattern grammar", |
| "authors": [ |
| { |
| "first": "Chung-Chi", |
| "middle": [], |
| "last": "Huang", |
| "suffix": "" |
| }, |
| { |
| "first": "Mei-Hua", |
| "middle": [], |
| "last": "Chen", |
| "suffix": "" |
| }, |
| { |
| "first": "Shih-Ting", |
| "middle": [], |
| "last": "Huang", |
| "suffix": "" |
| }, |
| { |
| "first": "Jason", |
| "middle": [], |
| "last": "Chang", |
| "suffix": "" |
| } |
| ], |
| "year": 2011, |
| "venue": "Proceedings of the 49th Annual Meeting of the Association for Computational Linguistics: Human Language Technologies", |
| "volume": "", |
| "issue": "", |
| "pages": "26--31", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Chung-Chi Huang, Mei-Hua Chen, Shih-Ting Huang, Jason Chang (2011). EdIt: A Broad-coverage Grammar Checker Using Pattern grammar. In Proceedings of the 49th Annual Meeting of the Association for Computational Linguistics: Human Language Technologies (p. 26-31).", |
| "links": null |
| }, |
| "BIBREF1": { |
| "ref_id": "b1", |
| "title": "Compounding Old and New Words in Filipino", |
| "authors": [ |
| { |
| "first": "Consuelo", |
| "middle": [ |
| "J" |
| ], |
| "last": "Paz", |
| "suffix": "" |
| } |
| ], |
| "year": 2003, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Consuelo J. Paz. (2003). Compounding Old and New Words in Filipino. University of the Philippines.", |
| "links": null |
| }, |
| "BIBREF2": { |
| "ref_id": "b2", |
| "title": "The Cambridge Learner Corpus -error coding and analysis for lexicography and ELT", |
| "authors": [ |
| { |
| "first": "Diane", |
| "middle": [], |
| "last": "Nicholls", |
| "suffix": "" |
| } |
| ], |
| "year": 1999, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Diane Nicholls. 1999. The Cambridge Learner Corpus -error coding and analysis for lexicography and ELT. Cambridge University Press.", |
| "links": null |
| }, |
| "BIBREF3": { |
| "ref_id": "b3", |
| "title": "The Tagalog Stemming Algorithm (TagSA)", |
| "authors": [ |
| { |
| "first": "Don", |
| "middle": [ |
| "Erick" |
| ], |
| "last": "Bonus", |
| "suffix": "" |
| } |
| ], |
| "year": 2003, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Don Erick Bonus. 2003. The Tagalog Stemming Algorithm (TagSA). Master's Thesis. De la Salle University, Manila.", |
| "links": null |
| }, |
| "BIBREF5": { |
| "ref_id": "b5", |
| "title": "A Sentence Syntax and Semantic Based Grammar Checker for Filipino", |
| "authors": [ |
| { |
| "first": "Panuring", |
| "middle": [], |
| "last": "Pampanitikan", |
| "suffix": "" |
| } |
| ], |
| "year": null, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Panuring Pampanitikan (PanPam): A Sentence Syntax and Semantic Based Grammar Checker for Filipino. Undergraduate Thesis. De La Salle University, Manila.", |
| "links": null |
| }, |
| "BIBREF6": { |
| "ref_id": "b6", |
| "title": "A Grammar Checker for Tagalog using LanguageTool", |
| "authors": [ |
| { |
| "first": "Nathaniel", |
| "middle": [], |
| "last": "Oco", |
| "suffix": "" |
| }, |
| { |
| "first": "Allan", |
| "middle": [], |
| "last": "Borra", |
| "suffix": "" |
| } |
| ], |
| "year": 2011, |
| "venue": "Proceedings of the 9th Workshop on Asian Language Resources Collocated with IJCNLP", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Nathaniel Oco and Allan Borra. 2011. A Grammar Checker for Tagalog using LanguageTool. Proceedings of the 9th Workshop on Asian Language Resources Collocated with IJCNLP 2011.", |
| "links": null |
| }, |
| "BIBREF7": { |
| "ref_id": "b7", |
| "title": "A Method for Unsupervised Broad-Coverage Lexical Error Detection and Correction", |
| "authors": [ |
| { |
| "first": "Nai-Lung", |
| "middle": [], |
| "last": "Tsao & David Wible", |
| "suffix": "" |
| } |
| ], |
| "year": 2009, |
| "venue": "Proceedings of the NAACL HLT Workshop on Innovative Use of NLP for Building Educational Applications", |
| "volume": "", |
| "issue": "", |
| "pages": "51--54", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Nai-Lung Tsao & David Wible. 2009. A Method for Unsupervised Broad-Coverage Lexical Error Detection and Correction. Proceedings of the NAACL HLT Workshop on Innovative Use of NLP for Building Educational Applications, p. 51-54.", |
| "links": null |
| }, |
| "BIBREF8": { |
| "ref_id": "b8", |
| "title": "Developing a chunk-based grammar checker for translated English sentences", |
| "authors": [ |
| { |
| "first": "", |
| "middle": [], |
| "last": "Nay Yee Lin", |
| "suffix": "" |
| } |
| ], |
| "year": 2011, |
| "venue": "Proceedings of PACLIC-2011", |
| "volume": "", |
| "issue": "", |
| "pages": "245--254", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Nay Yee Lin, Khin Mar Soe, and Ni Lar Thein. 2011. Developing a chunk-based grammar checker for translated English sentences. In Proceedings of PACLIC-2011 (p. 245-254).", |
| "links": null |
| }, |
| "BIBREF9": { |
| "ref_id": "b9", |
| "title": "Komisyon sa Wikang Pilipino", |
| "authors": [ |
| { |
| "first": "Ortograpiyang", |
| "middle": [], |
| "last": "Pambansa", |
| "suffix": "" |
| } |
| ], |
| "year": 2013, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Ortograpiyang Pambansa. 2013. Komisyon sa Wikang Pilipino.", |
| "links": null |
| }, |
| "BIBREF10": { |
| "ref_id": "b10", |
| "title": "Google Books N-gram Corpus used as a Grammar Checker", |
| "authors": [ |
| { |
| "first": "Rogelio", |
| "middle": [], |
| "last": "Nazar", |
| "suffix": "" |
| }, |
| { |
| "first": "& Irene", |
| "middle": [], |
| "last": "Renau", |
| "suffix": "" |
| } |
| ], |
| "year": 2012, |
| "venue": "Proceedings of the EACL 2012 Workshop on Computational Linguistics and Writing", |
| "volume": "", |
| "issue": "", |
| "pages": "27--34", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Rogelio Nazar & Irene Renau. 2012. Google Books N-gram Corpus used as a Grammar Checker. Proceedings of the EACL 2012 Workshop on Computational Linguistics and Writing, p. 27-34.", |
| "links": null |
| }, |
| "BIBREF11": { |
| "ref_id": "b11", |
| "title": "Makabagong Bararila ng Pilipino", |
| "authors": [ |
| { |
| "first": "Teresita", |
| "middle": [ |
| "V" |
| ], |
| "last": "Ramos", |
| "suffix": "" |
| } |
| ], |
| "year": 1971, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Teresita V. Ramos. 1971. Makabagong Bararila ng Pilipino. Rex Book Store.", |
| "links": null |
| }, |
| "BIBREF13": { |
| "ref_id": "b13", |
| "title": "TPOST: A Template-based Part-of-Speech Tagger for Tagalog", |
| "authors": [], |
| "year": null, |
| "venue": "Journal of Research in Science, Computing, and Engineering", |
| "volume": "3", |
| "issue": "1", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "TPOST: A Template-based Part-of-Speech Tagger for Tagalog. Journal of Research in Science, Computing, and Engineering, 3(1).", |
| "links": null |
| }, |
| "BIBREF14": { |
| "ref_id": "b14", |
| "title": "Manila: Lexmedia Digital Corporation", |
| "authors": [ |
| { |
| "first": "", |
| "middle": [], |
| "last": "Wikapedia", |
| "suffix": "" |
| } |
| ], |
| "year": 2015, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Wikapedia (2015), Manila: Lexmedia Digital Corporation.", |
| "links": null |
| } |
| }, |
| "ref_entries": { |
| "TABREF3": { |
| "content": "<table><tr><td colspan=\"2\">Correct</td><td/><td colspan=\"2\">Misspelled</td><td>Reason</td></tr><tr><td>Word</td><td/><td/><td>as</td><td/></tr><tr><td colspan=\"2\">Pangkain</td><td/><td colspan=\"2\">Pang kain</td><td>Extra Space</td></tr><tr><td>'used</td><td/><td>for</td><td/><td/></tr><tr><td>eating'</td><td/><td/><td/><td/></tr><tr><td colspan=\"2\">Tagtuyo</td><td/><td colspan=\"2\">Tag-tuyo</td><td>Extra Hyphen</td></tr><tr><td colspan=\"2\">'drought'</td><td/><td/><td/></tr><tr><td>Ikawalo</td><td/><td/><td colspan=\"2\">Ika-walo</td><td>Extra Hyphen</td></tr><tr><td>'eighth'</td><td/><td/><td/><td/></tr><tr><td colspan=\"3\">i-predict 'to</td><td>ipredict</td><td/><td>Missing</td></tr><tr><td>predict'</td><td/><td/><td/><td/><td>Hyphen</td></tr><tr><td>mas</td><td colspan=\"2\">malaki</td><td colspan=\"2\">masmalaki</td><td>Missing Space</td></tr><tr><td>'bigger'</td><td/><td/><td/><td/></tr><tr><td colspan=\"2\">inilagay</td><td>sa</td><td>nilagay</td><td>sa</td><td>Incorrect</td></tr><tr><td colspan=\"3\">kahon 'placed</td><td>kahon</td><td/><td>Affix used for</td></tr><tr><td colspan=\"2\">in a box'</td><td/><td/><td/><td>a verb focus</td></tr><tr><td/><td/><td colspan=\"4\">Table 3: Affix Errors</td></tr><tr><td colspan=\"4\">Confused between:</td><td/></tr><tr><td>ng 'of'</td><td/><td/><td/><td colspan=\"2\">nang '(function word</td></tr><tr><td/><td/><td/><td/><td colspan=\"2\">before an adverb)'</td></tr><tr><td colspan=\"4\">may 'has (used before</td><td colspan=\"2\">mayroon 'has (used</td></tr><tr><td colspan=\"3\">nouns, verbs,</td><td/><td colspan=\"2\">before grammatical</td></tr><tr><td colspan=\"3\">adjectives and</td><td/><td colspan=\"2\">particles, personal</td></tr><tr><td colspan=\"2\">adverbs)'</td><td/><td/><td colspan=\"2\">pronouns, and adverbs</td></tr><tr><td/><td/><td/><td/><td>of place'</td></tr><tr><td colspan=\"4\">suffix -ng 'used in</td><td colspan=\"2\">na '(type of</td></tr><tr><td colspan=\"4\">place of na if word</td><td colspan=\"2\">grammatical particle)'</td></tr><tr><td colspan=\"4\">preceding it ends in a</td><td/></tr><tr><td>vowel</td><td/><td/><td/><td/></tr></table>", |
| "num": null, |
| "html": null, |
| "text": "Spelling Errors", |
| "type_str": "table" |
| }, |
| "TABREF4": { |
| "content": "<table/>", |
| "num": null, |
| "html": null, |
| "text": "Wrong Word Usage", |
| "type_str": "table" |
| }, |
| "TABREF6": { |
| "content": "<table><tr><td colspan=\"4\">Error Type Weight Error Type Weight</td></tr><tr><td>Incorrect</td><td>0.6</td><td>Wrong</td><td>0.8</td></tr><tr><td>Affix/ Form</td><td/><td>Word Same</td><td/></tr><tr><td/><td/><td>POS</td><td/></tr><tr><td>Spelling</td><td>0.65</td><td>Wrong</td><td>0.95</td></tr><tr><td>Error</td><td/><td>Word Diff.</td><td/></tr><tr><td/><td/><td>POS</td><td/></tr><tr><td>Incorrectly</td><td>0.7</td><td>Missing</td><td>1.0</td></tr><tr><td>Merged</td><td/><td>Word</td><td/></tr><tr><td>Incorrectly</td><td>0.7</td><td>Unnecessary</td><td>1.0</td></tr><tr><td>Unmerged</td><td/><td>Word</td><td/></tr></table>", |
| "num": null, |
| "html": null, |
| "text": ".", |
| "type_str": "table" |
| }, |
| "TABREF7": { |
| "content": "<table/>", |
| "num": null, |
| "html": null, |
| "text": "Edit Distance Values", |
| "type_str": "table" |
| } |
| } |
| } |
| } |
