This study explores the complementarity of Machine Learning tools for performance analysis. We mainly focus on applying the supervised Machine Learning (ML) framework to improve the performance evaluation and data-driven goal setting of food processing companies in India (2011–21). Mathematical programming-based Data Envelopment Analysis (DEA) is the most popular tool for performance evaluation of decision-making units(DMUs) such as hospitals, companies, and local self-government institutions, etc. The relative benchmarking efficiency frontier is framed by using the given input(s)-output(s) of each DMUs. Disregarding the wider usage and applications in various fields, this tool is criticized for its sensitivity in the context of finite samples. In this study, we propose an Imputation Propelled DEA (IPDEA) as an alternative to overcome this limitation of conventional Non-Parametric DEA. In this study, we propose an Imputation Propelled DEA (IPDEA) using Machine learning algorithms to overcome the issues of the missing values in the data for the non-parametric Data Envelopment Analysis (DEA), and the prediction of the target variable. We consider Indian food processing industries to evaluate the proposed framework and presented the cases to demonstrate the proposed approach. In short, the DEA efficiency scores are sensitive to the sample size, and the proposed IPDEA can address the issue of finite samples to the best of its ability. Combining DEA with machine learning algorithms enhances our ability to accurately assess the Indian industries’ performance, which we can carry out through two parallel analyses, i.e., one with only actual data and the other with imputed data (along with actual data). Our results show a significant improvement in estimating efficiency scores using IPDEA.