Loading...

Generating a Quran’s Language-Based Arabic Text as The Most Relevant result to a Persian Text

Abdollahi Moghadam, Behdad | 2024

17 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 56948 (19)
  4. University: Sharif University of Technology
  5. Department: Computer Engineering
  6. Advisor(s): Izadi, Mohammad
  7. Abstract:
  8. Natural language processing has been used in solving the problems of humanities, including linguistics and analysis of religious books, such as the Bible, the Torah, and the Quran. One of the challenges in this field is the analysis of issues using the Quran. The problem in this issue is that a person becomes prejudiced in order to analyze the issues and makes incorrect judgements either inadvertently or intentionally based on his/her intellectual background. The same applies to analysis of issues based on the Quran. The method of analysis in this research is to produce an Arabic text based on the language of the Quran by receiving subjects in Persian. In order to achieve this goal, three approaches, namely English summarization, Arabic summarization, and text generation approaches are investigated and each of them is evaluated and their results are compared with each other. In addition, in each of these three approaches, there is a common step for information retrieval, which is used to obtain verses similar to the input text, and this step is done in two different methods. In the first method, an application programming interface that uses the semantic similarity between languages is used, and in the second method, which is implemented for this thesis, a monolingual semantic similarity is used, in this regard, the input text, which is in Persian, is converted into English and compared with the English equivalent of the verses. Again, the performance of these two methods is also compared with each other. The English summarization, after the information retrieval stage, has the steps of translating the Arabic text into English, summarizing the English text, and then translating the English text into Arabic, respectively. The Arabic summarization approach, after the information retrieval stage, has only the Arabic text summarization stage. The text generation approach is similar to the English summarization approach in all its steps, with this difference that in this approach, English text generation is used instead of the summarization step. Finally, we show that the performance of the Arabic summarization approach is slightly better than the English summarization approach and these two approaches are better than the text generation approach. Also, it is shown that among the two methods of finding verses similar to the input text, the method which is implemented for this thesis, has better performance than the method that uses the application programming interface
  9. Keywords:
  10. Natural Language Processing ; Deep Learning ; Text Analysis ; Quran’s Language

 Digital Object List

 Bookmark

No TOC