Axiomatic Ontology Learning Approaches for English Translation of the Meaning of Quranic Texts

Abstract

Abstract. Ontology learning (OL) is the computational task of generating a knowledge base in the form of an ontology, given an unstructured corpus in natural language (NL). While most works in the field of ontology learning have been primarily based on a statistical approach to extract lightweight OL, very few attempts have been made to extract axiomatic OL (called heavyweight OL) from NL text documents. Axiomatic OL supports more precise formal logic-based reasoning when compared to lightweight OL. Lexico-syntactic pattern matching and statisticsal one cannot lead to very accurate learning, mostly because of several linguistic nuances in the NL. Axiomatic OL is an alternative methodology that has not been explored much, where a deep linguistics analysis in computational linguistics is used to generate formal axioms and definitions instead of simply inducing a taxonomy. The ontology that is created not only stores the information about the application domain in explicit knowledge, but also can deduce the implicit knowledge from this ontology. This research will explore the English translation of the meaning of Quranic texts.

Extracting Knowledge From English Translated Quran Using NLP Pattern

“…To date, many ontology have been developed for Quran focusing on certain themes or domains [3], [7], [18]. Ontology for Quran as a whole, covering all domains in Quran is still not complete.…”

“…Ontology for Quran as a whole, covering all domains in Quran is still not complete. For example, the ontology for 'solat' only covers 10% from the entire Quran [3]. Ontology for representing semantic lexicons of Arabic about 'time' only focus on verses related to 'time' [7].…”

“…The selection of domains or verses to develop the ontology is based on the purposes of a project. Most of the construction were purposely done to support projects such as semantic search [19], [20], knowledge based [3], [18] and knowledge representation [7]. Other reason for selecting the verses are the difficulty exists that in understanding the meaning of each word in Quran.…”

“…Users need to relate a topic with few other verses that can be found in different chapters to get a whole picture of the topic [2]. This is due to the scattered knowledge in the Quran [3], [4] which are the main uniqueness of the Quran. Through the availability of Semantic Web technology, the scattered knowledge in Quran can be represented as ontology.…”

“…Although the term extraction is more or less stabilized with the performance above 90% in F-score [17] but this does not happen in Quran text translation. So far, research to discover the term extraction from Quran text translation performs 77% recall after enhancing the Natural Language Processing (NLP) pattern and ranking using statistical analysis [3]. This paper addresses the lack of current ontology development for Quran.…”