{ "paper_id": "2021", "header": { "generated_with": "S2ORC 1.0.0", "date_generated": "2023-01-19T07:33:25.693800Z" }, "title": "A 3 C: Arabic Anaphora Annotated Corpus", "authors": [ { "first": "Abdelhalim", "middle": [ "Hafedh" ], "last": "Dahou", "suffix": "", "affiliation": { "laboratory": "", "institution": "Ahmed Draia University Adrar", "location": { "country": "Algeria" } }, "email": "" }, { "first": "Mohamed", "middle": [], "last": "Abdelmoazz", "suffix": "", "affiliation": { "laboratory": "", "institution": "Ahmed Draia University Adrar", "location": { "country": "Algeria" } }, "email": "m.abdelmoazz@yahoo.com" }, { "first": "Mohamed", "middle": [ "Amine" ], "last": "Cheragui", "suffix": "", "affiliation": { "laboratory": "", "institution": "Ahmed Draia University Adrar", "location": { "country": "Algeria" } }, "email": "m_cheragui@univ-adrar.edu.dz" } ], "year": "", "venue": null, "identifiers": {}, "abstract": "In this paper, we describe the different steps taken to build our annotated corpus which aims to treat a known linguistic phenomenon in Arabic texts called Anaphora. The objective behind the creation of this corpus 1 is to fill the lack of resources concerning the resolution anaphora (especially pronominal and verbal) in the Modern Standard Arabic language and this is by creating a newly annotated corpus that we have called A 3 C which contains the anaphoric relations. To satisfy this objective, we created A 3 T, an anaphoric annotating tool that uses linguistic and statistical rules to automatically detect anaphors and their referents. After that, we resort to human specialists to verify and correct our A 3 T annotation's errors for the corpus's credibility. This study discusses novel features that can aid in determining the best reference, as well as the problem of the lack of resources for verbal anaphora. 2 Varieties of Anaphora in Arabic text What makes the anaphora resolution mechanism complex in natural language processing in general and in Arabic, in particular, is the fact that it can", "pdf_parse": { "paper_id": "2021", "_pdf_hash": "", "abstract": [ { "text": "In this paper, we describe the different steps taken to build our annotated corpus which aims to treat a known linguistic phenomenon in Arabic texts called Anaphora. The objective behind the creation of this corpus 1 is to fill the lack of resources concerning the resolution anaphora (especially pronominal and verbal) in the Modern Standard Arabic language and this is by creating a newly annotated corpus that we have called A 3 C which contains the anaphoric relations. To satisfy this objective, we created A 3 T, an anaphoric annotating tool that uses linguistic and statistical rules to automatically detect anaphors and their referents. After that, we resort to human specialists to verify and correct our A 3 T annotation's errors for the corpus's credibility. This study discusses novel features that can aid in determining the best reference, as well as the problem of the lack of resources for verbal anaphora. 2 Varieties of Anaphora in Arabic text What makes the anaphora resolution mechanism complex in natural language processing in general and in Arabic, in particular, is the fact that it can", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Abstract", "sec_num": null } ], "body_text": [ { "text": "A corpus is considered today as a fundamental piece in natural language processing, due to the role that it plays in both the resolution and the testing phases. The building of annotated corpus in terms of number and size has known a real ascension in the last decades, in particular since the appearance of statistical and machine learning approaches (Beseiso and Al-Alwani, 2016), allowing, from textual resources, the development of resolution models for different linguistic phenomena such as anaphora.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Introduction", "sec_num": "1" }, { "text": "Anaphora is typically defined as references to items mentioned earlier in a discourse or \"pointing back \" reference as described by (Mitkov, 99) . In addition, the process of determining the referent of an anaphora and establishing the relationship between them is known as anaphora resolution. Anaphora still a very challenging linguistic phenomenon, where its identification and resolution can increase the performance of several NLP applications, such as: sentiment analysis (Cambria, 2016) , question-answer systems (El-Said Nada et al., 2018) , machine translation (Madhura and Satish, 2019), text summarization (Antunes et al., 2018) , information extraction (Matysiak, 2007) , language generation and dialog systems (Vinay et al., 2019) .", "cite_spans": [ { "start": 132, "end": 140, "text": "(Mitkov,", "ref_id": null }, { "start": 141, "end": 144, "text": "99)", "ref_id": null }, { "start": 478, "end": 493, "text": "(Cambria, 2016)", "ref_id": "BIBREF5" }, { "start": 520, "end": 547, "text": "(El-Said Nada et al., 2018)", "ref_id": null }, { "start": 617, "end": 639, "text": "(Antunes et al., 2018)", "ref_id": "BIBREF1" }, { "start": 665, "end": 681, "text": "(Matysiak, 2007)", "ref_id": "BIBREF16" }, { "start": 723, "end": 743, "text": "(Vinay et al., 2019)", "ref_id": "BIBREF27" } ], "ref_spans": [], "eq_spans": [], "section": "Introduction", "sec_num": "1" }, { "text": "Our motivation behind this work is to enhance anaphora resolution in Arabic text by building an anaphoric annotated corpus that can contribute to future works that tackle anaphora in the Arabic language.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Introduction", "sec_num": "1" }, { "text": "This paper is structured in 6 sections. Section 2, describe the anaphoric typology in Arabic language. Section 3, gives an overview of existing anaphoric corpora (case of Arabic). Section 4, presents the challenges we face in Arabic anaphora resolution. Section 5, outlines the different phases of building of our A 3 C corpus. Section 6, some observations noted during the building process of our corpus. The last section gives a conclusion and future work. manifest in different forms (linguistic categories: lexical and grammatical), but also requires knowledge at different levels, as well as an \"understanding\" of the context. There are many varieties of anaphora in the Arabic text, we will only mention the most frequent ones.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Introduction", "sec_num": "1" }, { "text": "Verbal anaphora is used to describe or represent various movements or actions by using the verb (did -\u202b\u0641\u0639\u0644\u202c -) and the different conjugation variants to minimize writing and avoid repetition (Trabelsi et al., 2016; Hamouda, 2014) . ", "cite_spans": [ { "start": 191, "end": 214, "text": "(Trabelsi et al., 2016;", "ref_id": "BIBREF26" }, { "start": 215, "end": 229, "text": "Hamouda, 2014)", "ref_id": "BIBREF11" } ], "ref_spans": [], "eq_spans": [], "section": "Verbal anaphora", "sec_num": "2.1" }, { "text": "Lexical anaphora occurs when the referent is designated by definite descriptions representing the same concept (the anaphora), or concepts that are semantically close (Hammami, 2009) . Usually, this form of anaphora adds more information to the sentence and increases cohesion, and can take several forms (synonym, generalization / hypernymy, or specialization /hyponymy) (Seddik and Farghaly, 2014). ", "cite_spans": [ { "start": 167, "end": 182, "text": "(Hammami, 2009)", "ref_id": "BIBREF10" } ], "ref_spans": [], "eq_spans": [], "section": "Lexical anaphora", "sec_num": "2.2" }, { "text": "This type of anaphora is manifested by the introduction of lexical modifiers (e.g., \u202b\u0623\u062e\u0631\u202c / other, \u202b\u0623\u062e\u0631\u0649\u202c / one, \u202b)\u0648\u062d\u062f\u0629\u202c or comparative adjectives ( \u202b\u0645\u0646\u202c \u202b\u0623\u0643\u0628\u0631\u202c / greater than , \u202b\u0645\u0646\u202c \u202b\u0623\u062d\u0633\u0646\u202c / better than) (Hammami, 2009) . This variety of anaphora indicates a relation like: such as set-complement, similarity and comparison between the anaphora and the referent (Mahmoud Seddik and Farghaly, 2014). ", "cite_spans": [ { "start": 205, "end": 220, "text": "(Hammami, 2009)", "ref_id": "BIBREF10" } ], "ref_spans": [], "eq_spans": [], "section": "Comparative anaphora", "sec_num": "2.3" }, { "text": "Based on statistical studies done by (Hammami, 2009) it shows that the pronominal anaphora is the most frequent variant in Arabic texts. Pronouns form a special class of an\"aphora because of their empty semantic structures; they have a meaning independent of its referents and usually refer to names or noun phrases (Beseiso and Al-Alwani, 2016). However, not all pronouns are anaphoric.", "cite_spans": [ { "start": 37, "end": 52, "text": "(Hammami, 2009)", "ref_id": "BIBREF10" } ], "ref_spans": [], "eq_spans": [], "section": "Pronominal anaphora", "sec_num": "2.4" }, { "text": "Pronominal Anaphors can be divided into three categories, each category can be subdivided into subcategories according to several parameters, such as gender, number, etc.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Pronominal anaphora", "sec_num": "2.4" }, { "text": "3 rd personal pronouns ( \u202b\u0627\u0644\u063a\u0627\u0626\u0628\u202c \u202b:)\u0636\u0645\u0627\u0626\u0631\u202c In the Arabic, not all personal pronouns are anaphoric, so the 1 st person ( \u202b\u0646\u062d\u0646\u202c \u202b\u0648\u202c \u202b)\u0627\u0646\u0627\u202c and 2 nd person ( \u202b\u0627\u0646\u062a\u0645\u0627\u202c \u202b\u0623\u0646\u062a\u202c .., etc.) pronouns are not (they specify the communication partners and their meaning goes back to their specific uses), except the 3 rd person pronouns which have this characteristic. These pronouns can be subdivided into two categories: disjoint pronoun (Example: \u202b\u0647\u064a\u202c / she \u202b\u0647\u0648\u202c \u060c / he) and joint pronoun (Example: ", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Pronominal anaphora", "sec_num": "2.4" }, { "text": "\u202b\u0640\u0647\u202c \u060c \u202b\u0627\u202c \u060c \u202b\u0646\u202c ) (El-Said", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Pronominal anaphora", "sec_num": "2.4" }, { "text": "Translation: \"Soumaia has sewn her sister's wedding dress and she is very excited\". In some cases the pronouns \" \u202b\"\u0640\u0647\u202c and \" \u202b\"\u0640\u0647\u0627\u202c are not anaphoric since they are not interpreted as related to an expression (referent). In this case we will call them pleonastic pronouns. The use of relative pronouns is possible if the referent denotes a process or situation, and here the anaphora denotes some of these lexical meanings. They refer to persons, places or things that are close or distant, the table below illustrates this type of pronouns.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Disjoint Personal Pronoun Anaphora", "sec_num": null }, { "text": "Demonstrative pronouns ( \u202b\u0623\u0633\u0645\u202c \u202b\u0627\u0625\u0644\u0634\u0627\u0631\u0629\u202c \u202b\u0627\u0621\u202c ): They are linguistic elements that accompany a designation gesture in order to coordinate the attention of the interlocutors when they are speaking (Jarbou, 2018). Generally, demonstrative pronouns are cataphoric and in some cases they can be anaphoric and even deixis (Bouzid et al., 2014) . Demonstratives agree in person, gender and number with their referent. In addition, there are pronouns, which are considered demonstratives, and which designate time and place (Example: \u202b\u0647\u0630\u0627\u202c / this, \u202b\u0647\u0646\u0627\u202c / here). ", "cite_spans": [ { "start": 317, "end": 338, "text": "(Bouzid et al., 2014)", "ref_id": null } ], "ref_spans": [], "eq_spans": [], "section": "Disjoint Personal Pronoun Anaphora", "sec_num": null }, { "text": "For Arabic language, a considerable effort has been made concerning the anaphoric phenomenon during the last two decades, which is reflected by several studies aiming in their majority to solve the problem of the pronominal anaphora. The objective of this section is to present an overview of works dedicated to building annotated corpus (anaphora identification and referent determination).", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Related Work", "sec_num": "3" }, { "text": "AnATAr (Hammami, 2009) 18895 words 2722 pairs of anaphor /referent. ", "cite_spans": [ { "start": 7, "end": 22, "text": "(Hammami, 2009)", "ref_id": "BIBREF10" } ], "ref_spans": [], "eq_spans": [], "section": "Anaphoric Resolution Category", "sec_num": null }, { "text": "The aim of this section is to present the main factors, which affect anaphoric resolution.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Ambiguities and anaphoric resolution", "sec_num": "4" }, { "text": "Without diacritics marks, an Arabic text is extremely unclear (morphologically and grammatically). According to (Debili and Achour, 1998), 74% of Arabic words might potentially take several lexical diacritization, making it difficult to determine if the anaphoric phenomenon or referent is the case. ", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Ambiguities and lack of diacritics", "sec_num": "4.1" }, { "text": "Translation: \"I know the student who went to China\". ", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Relative Pronoun Anaphora", "sec_num": null }, { "text": "The Arabic script is characterized by the agglutination phenomena, which is explained by the fact of combining numbers of words in just one. Compared to French or English, an Arabic word can sometimes correspond to a full sentence (Bouzida and Zribi, 2020). ", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Agglutination phenomenon", "sec_num": "4.2" }, { "text": "Arabic is a nearly free-order language. This order causes artificial syntactic ambiguities, since the grammar should provide all the possible combination rules for reversing the order of words in the sentence. For anaphora resolution, this type of flexibility is a problem for referent localization (Beseiso and Al-Alwani, 2016; Fotiadou et al., 2020). ", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Syntactic flexibility (Words free order)", "sec_num": "4.3" }, { "text": "This difficulty occurs when the referent is ambiguous (due to the presence of two or more referents for the same anaphora). In this case, external knowledge of the context is necessary to identify the correct referent (Brunner et al., 2002 ).", "cite_spans": [ { "start": 218, "end": 239, "text": "(Brunner et al., 2002", "ref_id": "BIBREF2" } ], "ref_spans": [], "eq_spans": [], "section": "Ambiguity of the referent", "sec_num": "4.4" }, { "text": "2 Joint Personal Pronoun \u00ab \u202b\u0647\u202c \u00bb are anaphoric. ", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Ambiguity of the referent", "sec_num": "4.4" }, { "text": "This case occurs when the anaphora refers to something, which is not present in the sentence or text. The Qur'anic text is an example where this phenomenon persists (Seddik and Farghaly, 2011), so in the example below the pronominal anaphora ( \u202b\u0647\u0648\u202c / he) refers to ( \u202b\u0644\u0644\u0627\u202c /Allah) which is not present in the \"Aya\". The human through his knowledge and reasoning system can easily make the connection between the pronominal anaphora ( \u202b\u0647\u0648\u202c / he) and ( \u202b\u0644\u0644\u0627\u202c /Allah). However, for anaphoric resolution systems the task is complicated. ", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Hidden referent", "sec_num": "4.5" }, { "text": "As mentioned above, the main objective is to provide an annotated resource that can be used in the automatic Arabic anaphora resolving systems. We decided to create an operational tool with a friendly interface that would help computer scientists and linguists to develop such resources. In this section, we'll go over the steps involved in building our corpus A 3 C and annotating it with our A 3 T system. We thought about breaking down the creation of our work environment into three Translation: \"And to Him belongs that which reposes by night and by day, and He is the Hearing, the Knowing.\"( Surah Al-An'\u0101m -13) (03) phases: data collection, anaphora resolution system, corpus annotation and verification. Each phase consists of essential modules that take place to accomplish the phase's purpose. ", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Building the A3 C", "sec_num": "5" }, { "text": "Our purpose is to build a corpus of texts from different fields to cover two types of anaphora, pronominal and verbal anaphora. The texts in our corpus are taken from the Alriyadh newspaper 4 , a daily Arabic newspaper, and they are divided into five categories: culture, sports, politics, economy, and miscellaneous. The choice of those categories is made after an analysis of different categories of texts in terms of the number and diversity of anaphora types. On the other hand, the choice of this newspaper is due to the volume of information, good structure of articles and diversity of categories. To attend to this objective, we developed a crawler system that takes as an input the URL of the category page and the limited number of articles, then returns as an output a cleaned text file in (.txt) format.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Data collection", "sec_num": "5.1" }, { "text": "We all know how effort and time consuming it is to manually resolve anaphora and annotate a text corpus. As a result, we created the A3T (Arabic Anaphora Annotating Program), a tool that manages resolution and annotation in an automatic way, while also providing a user-friendly interface to modify the results. The resolution process was divided into two sub-modules:", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Co-reference Resolution", "sec_num": "5.2" }, { "text": "Data Preparation: To help us address the anaphora problem, the text corpus must go through three processes. The first step is to break each text file into sentences using a sentence splitter mechanism based on the punctuations. Secondly, organizing these sentences in a specific input structure to prepare them for the POS and morphological analysis (Figure 11 ). Finally, determine which grammatical category a given word belongs to and other morphological features such as gender, number, state, voice. The MADAMIDA tool was chosen for our purposes because of its 95.9% precision and high-quality word-level disambiguation as mentioned in (Pasha et al., 2014) . The word-level disambiguation functionality will help us in the identification of the attached pronouns. Anaphora Resolution System: A 3 T allows the expert to select text to automatically detect and resolve anaphora. Once selected, the following three steps are applied to detect and resolve the problem :", "cite_spans": [ { "start": 641, "end": 661, "text": "(Pasha et al., 2014)", "ref_id": "BIBREF19" } ], "ref_spans": [ { "start": 350, "end": 360, "text": "(Figure 11", "ref_id": "FIGREF0" } ], "eq_spans": [], "section": "Co-reference Resolution", "sec_num": "5.2" }, { "text": "\u2022 Anaphora identification: Anaphora is identified by referring to their grammatical code, which is based on the MADAMIRA tag set. The output here is a list of all anaphora in the text with additional information like Id, Name, Gender, Number, and Sentence number. For the pronominal anaphora, the process differs from one type to another, for example, the POS tagging for pronominal attached anaphora doesn't have a tag for gender, number, and person because the output is in the attached form, we should apply a split mechanism to place each of them in their proper tag as illustrated in (Figure 13 ). On the other hand, for verbal anaphora identification, we combine all of the elements used for pronominal anaphora identification, such as gender, number, and so on, with a new feature that will aid in the resolution which is the voice feature (active or passive form). Tables 4 and 5 illustrate the distribution of the various types of anaphora in our corpus after applying this process. Table 4 : Statistics about the A 3 C corpus (A).", "cite_spans": [], "ref_spans": [ { "start": 589, "end": 599, "text": "(Figure 13", "ref_id": "FIGREF0" }, { "start": 992, "end": 999, "text": "Table 4", "ref_id": null } ], "eq_spans": [], "section": "Co-reference Resolution", "sec_num": "5.2" }, { "text": "Economy 1455", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Category Verbal Anaphora", "sec_num": null }, { "text": "Education 1294", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Category Verbal Anaphora", "sec_num": null }, { "text": "Politics 924", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Category Verbal Anaphora", "sec_num": null }, { "text": "Miscellany 1932 Table 5 : Statistics about the A 3 C corpus (B).", "cite_spans": [], "ref_spans": [ { "start": 16, "end": 23, "text": "Table 5", "ref_id": null } ], "eq_spans": [], "section": "Sport 1370", "sec_num": null }, { "text": "5 Pro. Ana: Pronominal Anaphora Figure 13 : Example of pronominal anaphora identification", "cite_spans": [], "ref_spans": [ { "start": 32, "end": 41, "text": "Figure 13", "ref_id": "FIGREF0" } ], "eq_spans": [], "section": "Sport 1370", "sec_num": null }, { "text": "\u2022 Identification of referent candidates: Referents are chosen based on their POS (nouns, NPs and proper noun) and a specific search scope is adjusted based on some tests and previous research (Mitkov, 99). The search scope is still not fixed in the case of anaphora, but based on analysis, a high number of references occurs on the two previous sentences. In our case, we took two sentences before and as a special case for the demonstrative anaphora, we took the same number after. For the case of verbal anaphora, in the active form, we took two sentences after the verb and for the passive or unknown form; we took two sentences before the verb. The selection considers all of a candidate's features, including gender, number, voice, definiteness, and sentence number.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Sport 1370", "sec_num": null }, { "text": "\u2022 Anaphora resolving: The goal is to choose the most appropriate referents from among the most likely candidates for each anaphora. We used morphological filters to remove unsuitable candidates by comparing gender, number, and existing sentence (search scope). To find the suitable referent, we used a collection of preferential factors that favor certain candidates over others, as shown in Table 6 . Each rule has a score that is fixed after a series of experiments that took into account previous work (Abolohom and Omar, 2017) . Each candidate was given a score for each rule, and the one with the highest overall score was recommended as Pronominal Anaphora the best referent. We chose the one that came closest to overcoming the score similarity.", "cite_spans": [ { "start": 505, "end": 530, "text": "(Abolohom and Omar, 2017)", "ref_id": "BIBREF0" } ], "ref_spans": [ { "start": 392, "end": 399, "text": "Table 6", "ref_id": "TABREF10" } ], "eq_spans": [], "section": "Sport 1370", "sec_num": null }, { "text": "Description A score of 1 is given if an NP is definite and of 0 if not.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Linguistic rules Description", "sec_num": null }, { "text": "A score of 1 is assigned to the recency (nearest one) NP to the anaphora and 0 if not.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Recency", "sec_num": null }, { "text": "A score of 2 is assigned to NPs in the previous sentence or two sentences and further than those are given 0.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Referential Distance", "sec_num": null }, { "text": "A score of 1 is issued to the first NP of each sentence and 0 if not.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "First Noun Phrases", "sec_num": null }, { "text": "A score of 1 is issued to the existing NP in title and 0 if not", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "NPs in the title", "sec_num": null }, { "text": "Scores of 1 are given to an NP that has the same morpho-syntactic features as the anaphora and 0 if not.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Grammatical function", "sec_num": null }, { "text": "A score of 2 is assigned to the most frequent NP in text and 0 if not. ", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Frequency of NP in text", "sec_num": null }, { "text": "This phase aims to annotate the text document using the obtained information from the previous phase, which is a list of pairs of anaphora and their appropriate referent, along with features like Gender, Number, Type, and POS. We used our tool A 3 T to make the annotation process simpler and fast. The tool offers a user-friendly interface to linguistic experts, allowing them to check and, if possible, change the connections between anaphora and its referent, resulting in a reliable corpus that can be used in other studies.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Corpus annotation and verification", "sec_num": "5.3" }, { "text": "More specifically, the interface displays the annotated text in the center, while all of the couples anaphora/candidates are displayed on the right, with the system's chosen couple.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Corpus annotation and verification", "sec_num": "5.3" }, { "text": "In this case, all the expert has to do is check whether the anaphora tag's number of referent matches the correct one, if not, he may adjust the number of referent to the correct one from the other suggested couples or create a new one if the system doesn't find out the correct antecedent.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Corpus annotation and verification", "sec_num": "5.3" }, { "text": "In the final part, the tool will add automatically the following tags for the referent and the anaphora: the first will be marked with the tag. The remaining elements (anaphora) will be marked with . We also include the features listed above in each referent and anaphora tag. Finally, the A 3 T will generate an XML file that contains the text with anaphoric relationship tags as shown in Figure 15 . ", "cite_spans": [], "ref_spans": [ { "start": 411, "end": 420, "text": "Figure 15", "ref_id": "FIGREF0" } ], "eq_spans": [], "section": "Corpus annotation and verification", "sec_num": "5.3" }, { "text": "In our A 3 T system's testing, we used the \"AnATAr\" corpora for the evaluation for both anaphora phenomena. We used the standard accuracy metric to calculate the efficiency of the A 3 T and Table 7 presents the results obtained. We were unaware of any prior works on verbal anaphora, so we tested our work by taking the help of a linguistics specialist and utilizing the same corpora.", "cite_spans": [], "ref_spans": [ { "start": 190, "end": 197, "text": "Table 7", "ref_id": null } ], "eq_spans": [], "section": "Results and Discussion", "sec_num": "6" }, { "text": "Pronominal anaphora", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Anaphora Type Corpus Accuracy achieved", "sec_num": null }, { "text": "Verbal anaphora AnATAr 57.23% Table 7 : The result of anaphora resolution system A 3 T on AnATAr corpora", "cite_spans": [], "ref_spans": [ { "start": 30, "end": 37, "text": "Table 7", "ref_id": null } ], "eq_spans": [], "section": "83.19%", "sec_num": null }, { "text": "After analyzing the output of our system, particularly for the verbal anaphora, we found some factors that have influenced our findings. The first factor is the word disambiguation tool limitation that can't in some cases specify the correct meaning of a word such as \" \u202b\"\u0633\u0645\u0648\u0647\u202c which can act as verb (name; designate) and noun (Highness; grace) or \" \u202b\"\u0627\u062e\u0627\u0647\u202c (brother, fraternize). The second factor is the search scope, which could also lead to the best referent being excluded from the list of referents due to being out of scope. In the automatic resolution, the tool rid the references that span multiple sentences but we correct this issue in the expert verification part. The third factor is that the MADAMIRA tool can't recognize composed words like \" \u202b\u0627\u0644\u0639\u0631\u0628\u064a\u0629\u202c \u202b\u0645\u0635\u0631\u202c \u202b\"\u062c\u0645\u0647\u0648\u0631\u064a\u0629\u202c (Arab Republic of Egypt) or even compound proper names that always occur together like \" \u202b\u0645\u062d\u0645\u202c \u202b\u0635\u0627\u0644\u062d\u202c \u202b\u062f\u202c \" (Mohamed Salah). Finally, in some situations, the voice feature causes a faulty judgment when deciding if the better referent occurs before or after the verb anaphora.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "83.19%", "sec_num": null }, { "text": "Anaphora plays an important role in understanding text and making it coherent. At the same time, it is still a challenging task in the Arabic language due to the complexity of language, the low number of tools, and the lack of linguistic resources. Our present work will make a contribution in the field of linguistic resources for anaphora in the Arabic language and that by providing an annotated corpus that takes into consideration the pronominal and the verbal type. In terms of reducing effort and time consuming during the phase of resolution and annotating, we created A 3 T, a tool that uses linguistic concepts to identify this phenomenon. With the help of the expert, we are sure that the A 3 C will be very useful to use in terms of developing intelligence tools that tackle the Arabic anaphora. For the perspectives, our vision will concentrate on the amelioration of the verbal resolution mechanism by using state-of-the-art tools and methods in computational linguistics and at the same time increase the size of the A 3 C corpus.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Conclusion", "sec_num": "7" }, { "text": "The Corpus is available for the community in : https://dahouabdelhalim.github.io/Anaphora-Corpus/", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "", "sec_num": null }, { "text": "https://www.alriyadh.com/", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "", "sec_num": null } ], "back_matter": [], "bib_entries": { "BIBREF0": { "ref_id": "b0", "title": "A Computational Model for Resolving Arabic Anaphora using Linguistic Criteria", "authors": [ { "first": "Abolohom", "middle": [], "last": "Abdullatif", "suffix": "" }, { "first": "Omar", "middle": [], "last": "Nazlia", "suffix": "" } ], "year": 2017, "venue": "Indian Journal of Science and Technology", "volume": "10", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Abolohom Abdullatif and Omar Nazlia. 2017. A Computational Model for Resolving Arabic Anaphora using Linguistic Criteria. Indian Journal of Science and Technology. Volume 10. Issue 3.", "links": null }, "BIBREF1": { "ref_id": "b1", "title": "Automatic cohesive summarization with pronominal anaphora resolution", "authors": [ { "first": "Antunes", "middle": [], "last": "Jamilson", "suffix": "" }, { "first": "Dueire Lins", "middle": [], "last": "Rafael", "suffix": "" }, { "first": "Lima", "middle": [], "last": "Rinaldo", "suffix": "" }, { "first": "Oliveira", "middle": [], "last": "Hil\u00e1rio", "suffix": "" }, { "first": "Riss", "middle": [], "last": "Marcelo", "suffix": "" }, { "first": "Simske", "middle": [], "last": "Steven", "suffix": "" } ], "year": 2018, "venue": "Computer Speech & Language", "volume": "52", "issue": "", "pages": "141--164", "other_ids": {}, "num": null, "urls": [], "raw_text": "Antunes Jamilson , Dueire Lins Rafael , Lima Rinaldo , Oliveira Hil\u00e1rio , Riss Marcelo , Simske Steven. 2018. Automatic cohesive summarization with pronominal anaphora resolution. Computer Speech & Language. Volume 52, page (s) 141-164.", "links": null }, "BIBREF2": { "ref_id": "b2", "title": "Pronominal Anaphora Resolution in the KANTOO Multilingual Machine Translation System. Language Technologies Institute", "authors": [ { "first": "", "middle": [], "last": "Baker Kathryn", "suffix": "" }, { "first": "Mitamura", "middle": [], "last": "Brunner Annelen", "suffix": "" }, { "first": "Nyberg", "middle": [], "last": "Teruko", "suffix": "" }, { "first": "Svoboda", "middle": [], "last": "Eric", "suffix": "" }, { "first": "", "middle": [], "last": "Dave", "suffix": "" }, { "first": "Enrique", "middle": [], "last": "Torrejon", "suffix": "" } ], "year": 2002, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Baker Kathryn, Brunner Annelen, Mitamura Teruko, Nyberg Eric, Svoboda Dave and Torrejon, Enrique. 2002. Pronominal Anaphora Resolution in the KANTOO Multilingual Machine Translation System. Language Technologies Institute, Carnegie Mellon University.", "links": null }, "BIBREF3": { "ref_id": "b3", "title": "A Coreference Resolution Approach using Morphological Features in Arabic", "authors": [ { "first": "Beseiso", "middle": [], "last": "Majdi", "suffix": "" }, { "first": "Al-Alwani", "middle": [], "last": "Abdulkareem", "suffix": "" } ], "year": 2016, "venue": "International Journal of Advanced Computer Science and Applications", "volume": "7", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Beseiso Majdi and Al-Alwani Abdulkareem, 2016. A Coreference Resolution Approach using Morphological Features in Arabic. International Journal of Advanced Computer Science and Applications. Volume 7. Issue 10.", "links": null }, "BIBREF4": { "ref_id": "b4", "title": "A Generic approach for Pronominal Anaphora and Zero Anaphora resolution in Arabic language", "authors": [ { "first": "Zribi Chiraz Ben", "middle": [], "last": "Bouzida Saoussen Mathlouthi", "suffix": "" }, { "first": "", "middle": [], "last": "Othmane", "suffix": "" } ], "year": 2020, "venue": "proceeding of the 24th International Conference on Knowledge-Based and Intelligent Information & Engineering Systems", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Bouzida Saoussen Mathlouthi and Zribi Chiraz Ben Othmane. 2020. A Generic approach for Pronominal Anaphora and Zero Anaphora resolution in Arabic language. In proceeding of the 24th International Conference on Knowledge-Based and Intelligent Information & Engineering Systems.", "links": null }, "BIBREF5": { "ref_id": "b5", "title": "Affective computing and sentiment analysis", "authors": [ { "first": "Cambria", "middle": [], "last": "Erik", "suffix": "" } ], "year": 2016, "venue": "IEEE Intelligent Systems", "volume": "31", "issue": "2", "pages": "102--107", "other_ids": {}, "num": null, "urls": [], "raw_text": "Cambria Erik. 2016. Affective computing and sentiment analysis. IEEE Intelligent Systems. Volume 31(2). Page (s) 102-107, 2016.", "links": null }, "BIBREF6": { "ref_id": "b6", "title": "Voyellation automatique de l'arabe", "authors": [ { "first": "Debili", "middle": [], "last": "Fathi", "suffix": "" }, { "first": "Hadh\u00e9mi", "middle": [], "last": "Achour", "suffix": "" } ], "year": 1998, "venue": "Proceedings of the Workshop on Computational Approaches to Semitic Languages. Page (s)", "volume": "", "issue": "", "pages": "42--49", "other_ids": {}, "num": null, "urls": [], "raw_text": "Debili Fathi and Hadh\u00e9mi Achour. 1998. Voyellation automatique de l'arabe. In Proceedings of the Workshop on Computational Approaches to Semitic Languages. Page (s). 42-49.", "links": null }, "BIBREF7": { "ref_id": "b7", "title": "A Syntactic based Approach to Anaphora Resolution in Arabic", "authors": [], "year": 2018, "venue": "proceeding of The Eighteenth Conference on Language Engineering (ESOLEC'18)", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "El-Said Nada Aya Nabil Mostafa, Saad Sameh and Al- Ansary Abou El-Magd. 2018. A Syntactic based Approach to Anaphora Resolution in Arabic. In proceeding of The Eighteenth Conference on Language Engineering (ESOLEC'18).", "links": null }, "BIBREF8": { "ref_id": "b8", "title": "Anaphora resolution and word order across adulthood: Ageing effects on online listening comprehension. Glossa: a journal of general linguistics", "authors": [ { "first": "Fotiadou", "middle": [], "last": "Georgia", "suffix": "" }, { "first": "Ana", "middle": [], "last": "Mu\u00f1oz", "suffix": "" }, { "first": "Tsimpli", "middle": [], "last": "P\u00e9rez", "suffix": "" }, { "first": "Mari", "middle": [], "last": "Ianthi", "suffix": "" } ], "year": 2020, "venue": "", "volume": "5", "issue": "", "pages": "1--29", "other_ids": {}, "num": null, "urls": [], "raw_text": "Fotiadou Georgia, Mu\u00f1oz Ana P\u00e9rez and Tsimpli Ianthi Mari. 2020. Anaphora resolution and word order across adulthood: Ageing effects on online listening comprehension. Glossa: a journal of general linguistics. Volume 5. Issue 1. Page (s) 1-29.", "links": null }, "BIBREF9": { "ref_id": "b9", "title": "Caract\u00e9risation Formelle des Ellipses de la Langue Arabe et Processus de Recouvrement de la Langue Arabe", "authors": [ { "first": "Haddar", "middle": [], "last": "Kais", "suffix": "" } ], "year": 2000, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Haddar Kais.2000. Caract\u00e9risation Formelle des Ellipses de la Langue Arabe et Processus de Recouvrement de la Langue Arabe. PhD thesis, University of Tunis.", "links": null }, "BIBREF10": { "ref_id": "b10", "title": "Arabic Anaphora Resolution: Corpora Annotation with Coreferential Links", "authors": [ { "first": "Hammami", "middle": [], "last": "Souha", "suffix": "" }, { "first": "Belguith", "middle": [], "last": "Lamia", "suffix": "" }, { "first": "Ben", "middle": [], "last": "Hamadou Abdelmajid", "suffix": "" } ], "year": 2009, "venue": "The International Arab Journal of Information Technology", "volume": "6", "issue": "5", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Hammami Souha, Belguith Lamia, and Ben Hamadou Abdelmajid. 2009. Arabic Anaphora Resolution: Corpora Annotation with Coreferential Links. The International Arab Journal of Information Technology. Volume 6, No. 5.", "links": null }, "BIBREF11": { "ref_id": "b11", "title": "Anaphora Resolution for Arabic Machine Translation: A Case Study of Nafs", "authors": [ { "first": "Hamouda", "middle": [], "last": "Wafya", "suffix": "" } ], "year": 2014, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Hamouda Wafya. 2014. Anaphora Resolution for Arabic Machine Translation: A Case Study of Nafs. Ph.D. dissertation, University of Newcastle Upon Tyne.", "links": null }, "BIBREF12": { "ref_id": "b12", "title": "Time frame as a determinant of accessibility of anaphoric demonstratives in Classical Arabic", "authors": [ { "first": "Jarbou", "middle": [], "last": "Samer", "suffix": "" }, { "first": "Omar", "middle": [], "last": "", "suffix": "" } ], "year": 2018, "venue": "Topics in Linguistics", "volume": "19", "issue": "", "pages": "57--71", "other_ids": {}, "num": null, "urls": [], "raw_text": "Jarbou Samer Omar. 2018. Time frame as a determinant of accessibility of anaphoric demonstratives in Classical Arabic. Topics in Linguistics. Volume 19. Issue 2. Page (s) 57-71.", "links": null }, "BIBREF13": { "ref_id": "b13", "title": "Pronoun Resolution Task for Multilingual Machine Translation", "authors": [ { "first": "Madhura", "middle": [], "last": "Phadke", "suffix": "" }, { "first": "Satish", "middle": [], "last": "Devane", "suffix": "" } ], "year": 2019, "venue": "Proceeding of the 5th International Conference on Next Generation Computing Technologies", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Madhura Phadke, and Satish Devane. 2019. Pronoun Resolution Task for Multilingual Machine Translation. In Proceeding of the 5th International Conference on Next Generation Computing Technologies (NGCT-2019).", "links": null }, "BIBREF14": { "ref_id": "b14", "title": "How to combine salience factors for Arabic Pronoun Anaphora Resolution", "authors": [], "year": 2017, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Mathlouthi Bouzid Saoussen, Zribi Chiraz Ben Othmane and Trabelsi F\u00e9riel Ben Fraj. 2017. How to combine salience factors for Arabic Pronoun Anaphora Resolution. In the proceeding of the 4th", "links": null }, "BIBREF15": { "ref_id": "b15", "title": "ACS/IEEE International Conference on Computer Systems and Applications", "authors": [], "year": null, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "ACS/IEEE International Conference on Computer Systems and Applications.", "links": null }, "BIBREF16": { "ref_id": "b16", "title": "Information Extraction Systems and Nominal Anaphora Analysis Needs", "authors": [ { "first": "Matysiak", "middle": [], "last": "Ireneusz", "suffix": "" } ], "year": 2007, "venue": "Proceedings of the International Multiconference on Computer Science and Information Technology. Page (s)", "volume": "", "issue": "", "pages": "183--192", "other_ids": {}, "num": null, "urls": [], "raw_text": "Matysiak Ireneusz. 2007. Information Extraction Systems and Nominal Anaphora Analysis Needs. In Proceedings of the International Multiconference on Computer Science and Information Technology. Page (s) 183-192.", "links": null }, "BIBREF17": { "ref_id": "b17", "title": "Anaphora resolution: The state of the art", "authors": [ { "first": "Mitkov", "middle": [], "last": "Ruslan", "suffix": "" } ], "year": 1999, "venue": "Paper based on the COLING'98/ACL'98", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Mitkov Ruslan 1999. Anaphora resolution: The state of the art. In Paper based on the COLING'98/ACL'98", "links": null }, "BIBREF19": { "ref_id": "b19", "title": "Madamira: A fast, comprehensive tool for morphological analysis and disambiguation of Arabic", "authors": [ { "first": "Pasha", "middle": [], "last": "Arfath", "suffix": "" }, { "first": "Al-Badrashiny", "middle": [], "last": "Mohamed", "suffix": "" }, { "first": "Diab", "middle": [], "last": "Mona", "suffix": "" }, { "first": "Eskander", "middle": [], "last": "El Kholy Ahmed", "suffix": "" }, { "first": "Habash", "middle": [], "last": "Ramy", "suffix": "" }, { "first": "Pooleery", "middle": [], "last": "Nizar", "suffix": "" }, { "first": "Rambow", "middle": [], "last": "Manoj", "suffix": "" }, { "first": "Roth", "middle": [], "last": "Owen", "suffix": "" }, { "first": "", "middle": [], "last": "Ryan", "suffix": "" } ], "year": 2014, "venue": "the Proceeding of LREC. Page (s)", "volume": "", "issue": "", "pages": "1094--1101", "other_ids": {}, "num": null, "urls": [], "raw_text": "Pasha Arfath , Al-Badrashiny Mohamed , Diab Mona , El Kholy Ahmed , Eskander Ramy , Habash Nizar , Pooleery Manoj , Rambow Owen , Roth Ryan. 2014. Madamira: A fast, comprehensive tool for morphological analysis and disambiguation of Arabic. In the Proceeding of LREC. Page (s) 1094- 1101.", "links": null }, "BIBREF20": { "ref_id": "b20", "title": "Anaphora Resolution. Chapter of book", "authors": [ { "first": "Mahmoud", "middle": [], "last": "Seddik Khadiga", "suffix": "" }, { "first": "Farghaly", "middle": [], "last": "Ali", "suffix": "" } ], "year": 2014, "venue": "Theory and Applications of Natural Language Processing", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Seddik Khadiga Mahmoud and Farghaly Ali. 2014. Anaphora Resolution. Chapter of book, Theory and Applications of Natural Language Processing, Springer.", "links": null }, "BIBREF21": { "ref_id": "b21", "title": "Arabic Anaphora Resolution Using Holy Qur'an Text As Corpus", "authors": [ { "first": "Mahmoud", "middle": [], "last": "Seddik Khadiga", "suffix": "" }, { "first": "Farghaly", "middle": [], "last": "Ali", "suffix": "" } ], "year": 2011, "venue": "proceeding of Arabic Language Technology International Conference (ALTIC)", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Seddik Khadiga Mahmoud and Farghaly.Ali. 2011. Arabic Anaphora Resolution Using Holy Qur'an Text As Corpus. In proceeding of Arabic Language Technology International Conference (ALTIC).", "links": null }, "BIBREF22": { "ref_id": "b22", "title": "Arabic Anaphora Resolution: Corpus of the Holy Qur'an Annotated with Anaphoric Information", "authors": [ { "first": "Seddik", "middle": [], "last": "Khadiga", "suffix": "" }, { "first": "Farghaly", "middle": [], "last": "Ali", "suffix": "" }, { "first": "Fahmy Aly", "middle": [], "last": "Aly", "suffix": "" } ], "year": 2015, "venue": "International Journal of Computer Applications", "volume": "124", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Seddik Khadiga, Farghaly Ali and Fahmy Aly Aly. 2015. Arabic Anaphora Resolution: Corpus of the Holy Qur'an Annotated with Anaphoric Information. International Journal of Computer Applications. Volume 124. Issue 15.", "links": null }, "BIBREF23": { "ref_id": "b23", "title": "QurAna. Corpus of the Quran annotated with Pronominal Anaphora", "authors": [ { "first": "Sharaf", "middle": [], "last": "Abdul Baquee", "suffix": "" }, { "first": "Atwell", "middle": [], "last": "Eric", "suffix": "" } ], "year": 2012, "venue": "the Proceedings of the Eight International Conference on Language Resources and Evaluation (LREC'12)", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Sharaf Abdul baquee.,and Atwell Eric. 2012. QurAna. Corpus of the Quran annotated with Pronominal Anaphora. In the Proceedings of the Eight International Conference on Language Resources and Evaluation (LREC'12).", "links": null }, "BIBREF24": { "ref_id": "b24", "title": "Anaphora and Coreference Resolution: A Review. Information Fusion", "authors": [ { "first": "Sukthanker", "middle": [], "last": "Rhea", "suffix": "" }, { "first": "Poria", "middle": [], "last": "Soujanya", "suffix": "" }, { "first": "Cambria", "middle": [], "last": "Erik", "suffix": "" }, { "first": "Thirunavukarasu", "middle": [], "last": "Ramkumar", "suffix": "" } ], "year": 2020, "venue": "", "volume": "59", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Sukthanker Rhea, Poria Soujanya, Cambria Erik and Thirunavukarasu Ramkumar. 2020. Anaphora and Coreference Resolution: A Review. Information Fusion. Volume 59.", "links": null }, "BIBREF25": { "ref_id": "b25", "title": "Arabic Anaphora Resolution Using Markov Decision Process", "authors": [], "year": 2016, "venue": "the Proceeding 17 th International Conference on Intelligent Text Processing and Computational Linguistics", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Trabelsi F\u00e9riel Ben Fraj, Zribi Chiraz Ben Othmane and Mathlouthi Saoussen. 2016. Arabic Anaphora Resolution Using Markov Decision Process. In the Proceeding 17 th International Conference on Intelligent Text Processing and Computational Linguistics.", "links": null }, "BIBREF26": { "ref_id": "b26", "title": "A Novel Approach Based on Reinforcement Learning for Anaphora Resolution", "authors": [ { "first": "", "middle": [], "last": "Trabelsi F\u00e9riel Ben Fraj", "suffix": "" } ], "year": 2016, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Trabelsi f\u00e9riel ben fraj. 2016. A Novel Approach Based on Reinforcement Learning for Anaphora Resolution. In the proceeding of the 28th International Business Information Management conference.", "links": null }, "BIBREF27": { "ref_id": "b27", "title": "Anaphora Resolution in Dialogue Systems for South Asian Languages", "authors": [ { "first": "Vinay", "middle": [], "last": "Annam", "suffix": "" }, { "first": "Nikhil", "middle": [], "last": "Koditala", "suffix": "" }, { "first": "Radhika", "middle": [], "last": "Mamidi", "suffix": "" } ], "year": 2019, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": { "arXiv": [ "arXiv:1911.09994" ] }, "num": null, "urls": [], "raw_text": "Vinay Annam, Nikhil Koditala and Radhika Mamidi. 2019. Anaphora Resolution in Dialogue Systems for South Asian Languages. arXiv:1911.09994.", "links": null } }, "ref_entries": { "FIGREF0": { "num": null, "type_str": "figure", "text": "Example of verbal anaphora.", "uris": null }, "FIGREF1": { "num": null, "type_str": "figure", "text": "Example of lexical anaphora.", "uris": null }, "FIGREF2": { "num": null, "type_str": "figure", "text": "Example of comparative anaphora.", "uris": null }, "FIGREF3": { "num": null, "type_str": "figure", "text": "Example of disjoint personal pronoun anaphora.", "uris": null }, "FIGREF4": { "num": null, "type_str": "figure", "text": "Example of pleonastic pronoun.Relative pronouns ( \u202b\u0627\u0644\u0645\u0648\u0635\u0648\u0644\u0629\u202c \u202b:)\u0627\u0623\u0644\u0633\u0645\u0627\u0621\u202c Relative pronouns in Arabic have the characteristic of being always anaphoric, in addition they have only one possible referent(Trabelsi, 2016) and refer to the immediate nominal phrase mentioned before(Bouzid et al., 2014) which they agree in gender and number.", "uris": null }, "FIGREF5": { "num": null, "type_str": "figure", "text": "Example of relative anaphora.", "uris": null }, "FIGREF6": { "num": null, "type_str": "figure", "text": "Example of demonstrative anaphora.", "uris": null }, "FIGREF7": { "num": null, "type_str": "figure", "text": "Example of agglutination.", "uris": null }, "FIGREF8": { "num": null, "type_str": "figure", "text": "Example of ambiguity of the referent.", "uris": null }, "FIGREF9": { "num": null, "type_str": "figure", "text": "Example of hidden referent.", "uris": null }, "FIGREF10": { "num": null, "type_str": "figure", "text": "General architecture of building A 3 C.", "uris": null }, "FIGREF11": { "num": null, "type_str": "figure", "text": "POS by MADAMIRA.", "uris": null }, "FIGREF12": { "num": null, "type_str": "figure", "text": "Score similarity (example in pronominal case).", "uris": null }, "FIGREF13": { "num": null, "type_str": "figure", "text": "Anaphora resolution heuristics.", "uris": null }, "FIGREF14": { "num": null, "type_str": "figure", "text": "Score similarity (example in pronominal case)", "uris": null }, "TABREF0": { "text": "", "content": "
Translation: \"Students who are about
to graduate must complete their
administrative file, if they do not, they will not receive their certificates.\"Referent
Verbal Anaphora
Translation: \"Ibn Sina was born in
Bukhara (in present-day Uzbekistan),
one of the most famous writings of
the scientist The Canon of Medicine\".Referent
Nada et
al., 2018):
Lexical Anaphora
", "html": null, "num": null, "type_str": "table" }, "TABREF1": { "text": "", "content": "
Translation: \"Departure of Vessels
transporting more than 50,000 tons
of agricultural nutrients (urea) and
various other from the industrial
port of AL Jubail in the Kingdom to
the Port Sudan in Khartoum.\"
Comparative Anaphora
", "html": null, "num": null, "type_str": "table" }, "TABREF3": { "text": "Example of ambiguities due to the lack of diacritics.", "content": "", "html": null, "num": null, "type_str": "table" }, "TABREF4": { "text": ": \"I think it is in our interest to work hard\".", "content": "
Pleonastic
Pronoun
Translation: \"This book is for my brother Samir.\"WordWord + Diacritics \u202b\u064e\u0640\u0628\u064e\u202c \u202b\u0640\u062a\u0640\u202c \u064e \u202b\u0643\u0640\u202c \u202b\u0640\u0628\u202c \u064f \u202b\u0640\u062a\u0640\u202c \u064f \u202b\u0643\u0640\u202cTranslation he wrote books
\u202b\u0643\u0640\u0640\u062a\u0640\u0640\u0628\u202c\u202b\u0650\u0640\u0628\u064e\u202c \u202b\u0640\u062a\u0640\u202c \u064f \u202b\u0643\u0640\u202cWritten
\u202b\u0650\u0640\u0628\u064e\u202c \u202b\u0640\u062a\u0640\u202c \u064f \u202b\u0643\u0640\u202cwas caused to write
Referent
", "html": null, "num": null, "type_str": "table" }, "TABREF6": { "text": "Words free order in Arabic sentences.", "content": "", "html": null, "num": null, "type_str": "table" }, "TABREF7": { "text": "", "content": "
Translation:\"The
student successfully
passed the exam because he is assiduous, despite hisThe first Joint Personal Pronoun
difficulty\".refer to \"Student\".
Translation: \"and with their books\".Agglutinated wordThe second Joint Personal Pronoun refer to \"Exam\"
EncliticProclitic
(Their)(and with)
\u202b\u0647\u0645\u202c+\u202b\u0648\u0628\u0640\u202c
", "html": null, "num": null, "type_str": "table" }, "TABREF10": { "text": "The linguistic preferences and their respective Scores.", "content": "", "html": null, "num": null, "type_str": "table" } } } }