Exploring electric vehicle adoption in Indonesia using zero-shot aspect-based sentiment analysis

Abstract

Transportation emissions significantly contribute to greenhouse gases (GHG) and climate change. Electric vehicles (EVs) offer a promising solution to this problem. Despite the noteworthy adoption of EVs in emerging economies like China and Europe, the pace of EV rollout in Indonesia remains sluggish. Currently, the country's rate of EV adoption is below 0.3 %, leading to stagnation in the Indonesian electric vehicle market. Several factors have impeded the adoption of EVs in Indonesia. Previous studies have investigated consumer acceptance of EV adoption in specific countries using sentiment analysis. Various data analytics and machine learning techniques have been implemented in those studies. However, those studies predominantly rely on the traditional sentiment analysis method, which assigns a single sentiment classification to each document or sentence. On the other hand, this study aims to investigate consumer acceptance of electric vehicle adoption in Indonesia through Twitter conversations. It utilises the Zero-Shot Aspect-Based Sentiment Analysis (ABSA) approach, which can provide a more fine-grained analysis of the aspects discussed and their corresponding sentiments. The findings demonstrate the effectiveness of this approach in identifying discussed aspects within tweets, though its sentiment classification performance is limited. The research also uncovers crucial aspects of electric vehicle adoption and public sentiments on Twitter. These insights could provide valuable guidance to the government and other stakeholders regarding the concerns associated with EV adoption in Indonesia.