Exploring Sacred Texts: Leveraging Computer Science for Dataset Similarity Analysis in Religious Studies

Abstract

Studying the Quran and the Hadith side by side can help us understand that the two are fundamental and two main resources and essential wellspring of Islamic knowledge and law. There are many debates about similarities between those holy scriptures from many famous preachers and scholars. Technology can be used as an alternative solution to solve these problems. There are at least two overall approaches to determine text-similarity; the vector space model and semantic similarity —define the similarity or the distance. The similarity between words is often represented by a similarity between concepts associated with the words. This paper presents a method for identifying semantic sentence similarity among each sentence from each dataset using semantic relation of word senses between different synsets using WordNet path similarity and Wu-Palmer similarity. This method is also evaluated and has acceptable accuracy. Although both Path Similarity and Wu-Palmer Similarity successfully identify the similarity between two sentences; still, they have slightly different accuracy. The Wu-Palmer similarity is superior to path similarity when identifying sentences between Quran Sahih International and An-Nawawi Forty Hadith Translation. Looking ahead, we might be able to improve our results by using multipliers such as reverse document frequency (TF-IDF), combining the results of several steps in WordNet similarity, using vector space models, and optimal matching methods.