Indonesian E-Commerce Social Media Sentiment Dataset

The dataset contains Indonesian social media data on e-commerce, and is intended for research on sentiment analysis. The data set was prepared and labeled into three sentiment classes as Positive, Neutral and Negative. The dataset has been used in the paper Public Sentiment Analysis of Social Media Data on E-Commerce: A Comparative Study of Classical Machine Learning and Transformer-Based Models. It is used for comparative evaluation of classical machine learning and transformer-based approaches for Indonesian sentiment classification. The dataset contains preprocessed text data with sentiment labels for training and evaluation of the model. The preprocessing steps are text cleaning, text normalization and label standardization. The labeled data were verified by an annotation process to enhance the consistency of labeling. This dataset is provided for the purpose of facilitating research reproducibility, model evaluation, and further research in the area of Indonesian language sentiment analysis in the e-commerce domain.