Revealing Indonesian Netizens’ Sentiments Towards Halal Food: A Machine Learning Approach

Social media has become a prevalent platform for communication and information dissemination in various domains. In particular, the exponential surge in user-generated content on social media has opened up new possibilities for opinion mining, making it a valuable tool for analyzing consumer behaviour. Understanding public sentiments towards halal food is of paramount importance for stakeholders and decision-makers in the food industry. This paper presents an investigation into the sentiments of social media users in Indonesia concerning halal food-related. We obtained a halal-food dataset consisting of 123,385 tweets after data preprocessing. To extract meaningful features from textual data, we explore two vectorization techniques, namely CountVectorizer and TfidfVectorizer. Leveraging these feature representations, we employ various machine learning algorithms, including Naive Bayes, Logistic Regression, Random Forest, Support Vector Machine, AdaBoost, XGBoost, Multi-Layer Perceptron, and Convolutional Neural Network models. Our performance evaluation is based on key metrics like accuracy, precision, recall, and F1-score. The results show that the XGBoost model with CountVectorizer in sentiment analysis produces the highest accuracy with a value of 0.8269. Several topics were derived from both sentiments based on the frequency of words generated, including the halal industry, emotions and feelings, food ingredients and menus.