Nonparametric Estimation of the Production Frontier Using a Data-Fitting Technique

We propose a novel nonparametric approach for estimating the production frontier based on data-fitting technique. The proposed allows stochastic noise and provides decision-makers with general flexible estimation procedure to support facilitate decision-making process in context. A major feature of is that completely easy implement. Similar other existing approaches, our results an estimate piece-wise linear frontier. In contrast ones, evaluation each data point performed within unit-specific range. also naive method determining ranges point. performance examined using various simulated scenarios. For scenario, we compare methods, including envelopment analysis (DEA), (StoNED), (SFA). simulation suggest performs better than methods single input output case. Our can be easily extended multi-input setting. Moreover, shows its flexibility usefulness examples.