Hesitation Analysis with Kullback Leibler Divergence and Its Calculation on Temporal Data

Hesitation analysis plays a crucial role in decision-making processes by capturing the intermediary position between supportive and opposing information. This study introduces a refined approach to addressing uncertainty in decision-making, employing existing measures used in decision problems. Building on information theory, the Kullback–Leibler (KL) divergence is extended to incorporate additional insights, specifically by applying temporal data, as illustrated by time series data from two datasets (e.g., affirmative and dissent information). Cumulative hesitation provides quantifiable insights into the decision-making process. Accordingly, a modified KL divergence, which incorporates historical trends, is proposed, enabling dynamic updates using conditional probability. The efficacy of this enhanced KL divergence is validated through a case study predicting Korean election outcomes. Immediate and historical data are processed using direct hesitation calculations and accumulated temporal information. The computational example demonstrates that the proposed KL divergence yields favorable results compared to existing methods.