Analyzing Public Perception using Aspect Based Sentiment Analysis: Case Study of Capital Relocation Planning of Indonesia

The use of machine learning to analyze public perception has become a prominent theme in governance services, particularly in relation to Indonesia’s capital relocation planning. This focus arises due to the substantial budget allocated for the development of the new capital (IKN) from the state expenditure. Consequently, various public responses have emerged regarding this phenomenon, prompting an analysis to understand public perception of the capital city’s relocation. This study will employ machine learning technologies for sentiment analysis, specifically focusing on Aspect-Based Sentiment Analysis (ABSA) using public responses about the IKN relocation obtained directly from the social media platform X. ABSA is different with a traditional sentiment analysis method cause focused on each aspect and context. The proposed method of this research involves using Support Vector Machine (SVM) and Decision Tree algorithms to identify each aspect. The results of this study indicate that the SVM algorithm achieved an optimal testing accuracy of $85.78 \%$, with the majority of public discussion showing support for the IKN project. The sentiment analysis revealed several points conveyed by the public, including the perception that the development of the IKN project is necessary given the difficult economic conditions and its impact on the people and the environment.