Sentiment Analysis of Spiritual Teachings: Ensemble Learning Applied to Twitter Religious Text and Bhagavad Gita

Religious texts serve as guiding principles that shape moral values, ethics, and social conduct across cultures. It also provides profound insights into spirituality and human behavior. Despite its historical and cultural significance, the computational analysis of the text remains an under-explored area of research. State-of-the-art models have shown lower performance, partly due to assumptions of feature independence. Religious texts often contain metaphors and context-specific meanings, making such assumptions ineffective for semantic and sentiment analysis, ultimately leading to reduced accuracy. This study addresses this gap by conducting a sentiment analysis of the Twitter religious text using advanced machine-learning techniques. The dataset is annotated using the VADER (Valence Aware Dictionary and sEntiment Reasoner) lexicon, with features extracted through the Term Frequency-Inverse Document Frequency (TF-IDF). Unlike prior studies that predominantly rely on individual models, this work addresses their limitations by proposing an advanced ensemble approach. A robust voting classifier is developed, combining multiple machine-learning algorithms to achieve an accuracy of 95%. This ensemble model demonstrates superior performance in capturing the nuances of textual sentiment com-pared to standalone methods. This is further explored through the Bhagavad Gita text to evaluate the model’s effectiveness and enrich the research work. One limitation of this approach is that large ensemble models may face scalability challenges when handling extensive religious corpora. Additionally, the black-box nature of ensemble models makes it difficult to interpret how a specific sentiment is assigned, which could lead to unintended bias or misrepresentation of religious sentiments, raising potential ethical concerns. By integrating computational techniques with literary analysis, this research paves the way for future exploration of religious and philosophical texts using deep learning and semantic analysis. The proposed methodologies can be extended to other sacred texts, fostering interdisciplinary advancements at the intersection of artificial intelligence and the humanities.