Classification Model for Assessing Micro and Small Enterprises’ Awareness Level Towards Halal Certification Adoption

To become the global halal hub, the Indonesian government is accelerating Micro and Small Enterprises (MSE) to adopt halal compliance by offering incentives for certification submission. Selection of MSEs that receive the incentives should be done due to the limitations of the budget. The aim of this study is to develop a classification model to measure the level of awareness among MSEs regarding halal certification adoption that can give suggestions in selecting MSEs. The model was constructed by machine learning algorithms applied to survey data acquired from 243 MSE owners. The data includes demographic information and MSEs’ opinions toward halal compliance. Ten algorithms, including five single and five ensemble models were explored in order to get the best predictive classification model. Imbalanced data handling and hyperparameter tuning were applied to improve the predictive performance. The study shows that the Gradient Boosting and Adaptive Boosting models outperform other algorithms, with the best accuracy of 90.28%.