Comparative efficacy of machine-learning models in prediction of reducing uncertainties in biosurfactant production.
Journal:
Bioprocess and biosystems engineering
PMID:
31280373
Abstract
An accurate and reliable forecast of biosurfactant production with minimum error is useful in any bioprocess engineering. Bacterial isolate FKOD36 capable of producing biosurfactant was isolated in this study and pre-inoculums was prepared from the agar slants in a small test tube and incubated at 30 °C for 24 h at 120 rpm. Due to inherent non-linearity characteristics of the data set in a bioprocess, conventional modeling techniques are not adequate for predicting biosurfactant production in a microbiological process. The main contribution of the study was to compare two soft-computing models, i.e., support vector regression (SVR) and support vector regression coupled with firefly algorithm to evaluate the best performance of the two mentioned models. Based on the results it was noted that support vector regression coupled with firefly algorithm performs better compared to the simple SVR.