Most traditional approaches to anaphora resolution rely heavily on linguistic and domain knowledge. One of the disadvantages of developing a knowledge-based system, however, is that it is a very labour-intensive and time-consuming task. This paper presents a robust, knowledge-poor approach to resolving pronouns in technical manuals, which operates on texts pre-processed by a part-of-speech tagger. Input is checked against agreement and for a number of antecedent indicators. Candidates are assigned scores by each indicator and the candidate with the highest score is returned as the antecedent. Evaluation reports a success rate of 89.7% which is better than the success rates of the approaches selected for comparison and tested on the same data. In addition, preliminary experiments show that the approach can be successfully adapted for other languages with minimum modifications.

Robust pronoun resolution with limited knowledge

The nature of the Quran and its translations as classic Arabic and English texts reduces the accuracy of ordinary natural language processing tools such as pronominal anaphora resolution systems. Pronominal anaphora resolution simply involves finding an antecedent for anaphoric pronouns as the referring expressions of discourse. The performance of a pronominal anaphora resolution system is vitally related to the efficiency of pre-processing tools that analyze and prepare the input data for feeding the resolution algorithm. This paper proposes a novel pre-processing approach for pronoun extraction and pronoun mapping in the pronominal anaphora resolution system of English translations of the Quran, which facilitates the anaphora resolution, specifically for the English pronouns without an explicit antecedent that contributes close to 50% of the anaphoric relations in the Quran. This approach uses the morphologic, statistic and anaphoric knowledge that is extracted from the Arabic corpus of the Quran. For evaluating the arrangement, 1% of an English translation was annotated with labeling for all anaphoric and non-anaphoric English pronouns. These pronouns were aligned to the equivalent Arabic pronouns and linked to the concepts in the Arabic text. Through statistical results, it was shown that our rule-based pre-processing tools perform well. The precision, recall, and accuracy of pronoun extraction stage are 96.38%, 100%, and 99.5%, respectively. The result of mapping algorithm is promising whereby we score 85.51% in precision, 96.32% in recall, and 82.81% in accuracy.

A Rule-Based Approach For Pronoun Extraction And Pronoun Mapping In Pronominal Anaphora Resolution Of Quran English Translations

Anaphora resolution has become a major issue in natural language processing systems; in this work we propose a resolution approach in which texts are parsed by a definite clause grammar and then converted into an XML-tagged representation, where sentence elements are marked with discourse, syntactic, and semantic attributes. This extension was made primarily to test the viability of using XML tagged documents for anaphora resolution. The XML representation allows valuable text’s enrichment with anaphoric information in an elegant and easy way. The system’s performance arises primarily from the integration of multiple knowledge sources in a modular architecture and uses constraints and preferences to select the antecedent. The developed system proposes to resolve pronominal anaphora, namely personal pronouns for French language texts.

Engineering & Information Technology

Pronominal Anaphora Resolution Using XML Tagged Documents
Anaphora resolution has become a major issue in natural language processing systems; in this work we propose a resolution approach in which texts are parsed by a definite clause grammar and then converted into an XML-tagged representation, where sentence elements are marked with discourse, syntactic, and semantic attributes. This extension was made primarily to test the viability of using XML tagged documents for anaphora resolution. The XML representation allows valuable text’s enrichment with anaphoric information in an elegant and easy way. The system’s performance arises primarily from the integration of multiple knowledge sources in a modular architecture and uses constraints and preferences to select the antecedent. The developed system proposes to resolve pronominal anaphora, namely personal pronouns for French language texts.

Allaoua Refoufi

Papers

Pronominal Anaphora Resolution Using XML TaggedDocuments