Enhancement of Kansei Model for Political Security Threat Prediction Using Bi-LSTM

Online platforms serve as valuable sources for monitoring public sentiment related to political security threats. Thus, there is a crucial need to study the sentiment analysis and how it can be utilized to predict the threats. This study introduces a new approach to improve the political security threat prediction by proposing enhancement of Kansei model using deep learning Bidirectional Long Short-Term Memory (Bi-LSTM). Data from various sources, including social media and user comments was utilized to perform the prediction. This study discussed that utilization of manual methods to perform Kansei analysis from establishment of Kansei checklist to determine Kansei words can be improved by integrating Bi-LSTM to significantly enhanced the analysis process and predictive capabilities. Additionally, ADAM algorithm was adopted as optimizer. Experimental analysis was performed, and the results show that the enhanced Kansei model with Bi-LSTM and ADAM optimizer demonstrates improvement in accuracy and performance. This approach has the potential to assist authorities and organizations in making more informed and proactive decisions in addressing political security issues, thereby improving national stability and security.