Efficiency analysis of halal certification bodies in Indonesia: a hybrid data envelopment analysis and machine learning approach

This study aims to analyze Halal Certification Body (Lembaga Pemeriksa Halal (LPH)) efficiency in Indonesia by using the Data Envelopment Analysis (DEA) and Machine Learning (ML) methods. As the largest Muslim country in the world, with more than 200 million Muslims or approximately 12% of the world’s total Muslim population, Indonesia aims to become a global halal hub by 2024. Therefore, LPH efficiency measurement plays an important role in supporting this mission. The efficiency analysis is based on input and output factors, which are selected based on the characteristics of an LPH. An artificial neural network (ANN) and logistic regression (Logit) are used to classify LPHs into efficient and inefficient groups based on their DEA scores, as these two methods can complement DEA. The data used in this study are collected from LPHs and the Ministry of Religious Affairs in Indonesia. The results show that the DEA and ML methods can be used to identify efficient and inefficient LPHs and the factors affecting their efficiency. The DEA-ANN predictive model has an accuracy of 85.06%, which is higher than that of the DEA-logit model (79%). This study contributes to a more efficient and effective halal certification system in Indonesia, supporting the growth of the halal industry in this country.