﻿<?xml version="1.0" encoding="utf-8"?>
<ArticleSet>
  <ARTICLE>
    <Journal>
      <PublisherName>مرکز منطقه ای اطلاع رسانی علوم و فناوری</PublisherName>
      <JournalTitle>Journal of Information Systems and Telecommunication (JIST) </JournalTitle>
      <ISSN>2322-1437</ISSN>
      <Volume>14</Volume>
      <Issue>55</Issue>
      <PubDate PubStatus="epublish">
        <Year>2026</Year>
        <Month>11</Month>
        <Day>9</Day>
      </PubDate>
    </Journal>
    <ArticleTitle>A Hybrid Entity-Centric Approach to Persian Pronoun Resolution with BERT Enhancement</ArticleTitle>
    <VernacularTitle>A Hybrid Entity-Centric Approach to Persian Pronoun Resolution with BERT Enhancement</VernacularTitle>
    <FirstPage>245</FirstPage>
    <LastPage>267</LastPage>
    <ELocationID EIdType="doi">10.66224/jist.49358.14.55.245</ELocationID>
    <Language>en</Language>
    <AuthorList>
      <Author>
        <FirstName>Hassan</FirstName>
        <LastName>Haji Mohammadi</LastName>
        <Affiliation>شهید بهشتی</Affiliation>
      </Author>
      <Author>
        <FirstName>Alireza</FirstName>
        <LastName>Talebpour</LastName>
        <Affiliation>shahid Beheshti</Affiliation>
      </Author>
      <Author>
        <FirstName>Ahmad</FirstName>
        <LastName> Mahmoudi-Aznaveh</LastName>
        <Affiliation>shahid beheshti</Affiliation>
      </Author>
      <Author>
        <FirstName>Samaneh</FirstName>
        <LastName>Yazdani</LastName>
        <Affiliation>دانشگاه آزاد اسلامی</Affiliation>
      </Author>
    </AuthorList>
    <History PubStatus="received">
      <Year>2025</Year>
      <Month>1</Month>
      <Day>31</Day>
    </History>
    <Abstract>&lt;p&gt;Pronoun resolution, a critical aspect of coreference resolution in natural language processing, is essential for accurate text understanding and generation. This paper introduces a hybrid Persian pronoun resolution model that leverages rule-based and machine-learning approaches. Seven high-precision rule-based modules, tailored explicitly for Persian linguistic nuances, are combined with a random forest classifier enhanced with BERT contextual embeddings and cluster-level features. This approach effectively addresses the challenges of pronoun resolution, surpassing the performance of existing Persian systems on two corpora: the newly developed Mehr corpus (400 documents) and the established RCDAT corpus. Integrating BERT embeddings and cluster-level features significantly contributes to the model's superior performance, achieving a 4.68 F1-score improvement on Mehr and a 4.01 F1-score improvement on RCDAT. The proposed hybrid model demonstrates a practical and interpretable solution, contributing to Persian natural language processing research and applications.&lt;/p&gt;</Abstract>
    <ObjectList>
      <Object Type="Keyword">
        <Param Name="Value">Coreference Resolution</Param>
      </Object>
      <Object Type="Keyword">
        <Param Name="Value">Natural Language Processing</Param>
      </Object>
      <Object Type="Keyword">
        <Param Name="Value">Pronoun Resolution</Param>
      </Object>
      <Object Type="Keyword">
        <Param Name="Value">Corpus Development</Param>
      </Object>
      <Object Type="Keyword">
        <Param Name="Value">BERT</Param>
      </Object>
      <Object Type="Keyword">
        <Param Name="Value">Transformer Models</Param>
      </Object>
    </ObjectList>
    <ArchiveCopySource DocType="Pdf">http://jist.ir/en/Article/Download/49358</ArchiveCopySource>
  </ARTICLE>
</ArticleSet>