Nonlinear PESTLE Framework-Based Sentiment Analysis Using Machine Learning for Root Cause Identification

Abstract

Sentiment analysis has gained significance in understanding the intricate dynamics of public sentiment in contemporary social media and news communication. This research proposes a nonlinear sentiment analysis model that integrates the PESTLE framework with machine learning techniques to identify and quantify the influence of external factors—political, economic, social, technological, legal, and environmental—on sentiment patterns. The model employs a Support Vector Machine (SVM) classifier with nonlinear kernel transformations to capture complex relationships between sentiment triggers and influencing factors. Experimental results demonstrate the efficacy of the nonlinear approach in correlating sentiment classifications with their intrinsic influences, offering a more robust understanding of root causes. Future work will explore advanced nonlinear classifiers and additional external factors for enhanced real-world applicability.