An Explainable Natural Language Processing Framework for Semantic Analysis and Knowledge Retrieval from Islamic and Feminist Text Collections

Abstract

As religious, cultural, and social text collections become increasingly digitized, sophisticated computational techniques are needed to extract meaningful information without losing transparency, context, or interpretation. Digital Islamic and feminist text collections have complex linguistic elements, historical contexts, multiple interpretations, and socially sensitive concepts that cannot be adequately processed using traditional keyword-based retrieval systems and non-explainable NLP approaches. The paper proposes an explainable NLP conceptual framework that combines semantic analysis, explainability mechanisms, and knowledge retrieval approaches for culturally sensitive texts. The study uses a qualitative conceptual approach that involves the analysis of 21 academic articles on topics such as NLP, Explainable Artificial Intelligence (XAI), semantic models, information retrieval, Islamic text processing, feminist discourse analysis, and ethical AI. There are five main conceptual themes identified in the qualitative analysis. The designed model is built on five interrelated layers: text collection and processing, natural language processing methods, semantic analysis and representation, explainability techniques, and knowledge extraction and application. In the review, such vital problems like context vagueness, language differences, multi-meaning, implicitness, and biases in analyzing sensitive texts are underlined. Besides that, the conceptual analysis of the feminist text analysis issues shows some major thematic areas like cultural and historical difference, multiple theoretical standpoints, implicitness, evolution of the vocabulary, and NLP bias. These thematic areas are qualitative aspects derived from the analyzed literature and not any empirical statistical data. The result is an understandable and person-oriented Explainable NLP model that ensures reliable knowledge discovery and provides the basis for further digital humanities research.