作者: EFSTATHIOS Z. PANAGOU , CHRYSOULA C. TASSOU , ELEFTHERIOS K. A. SARAVANOS , GEORGE-JOHN E. NYCHAS
DOI: 10.4315/0362-028X-70.8.1909
关键词:
摘要: The growth profile of five strains lactic acid bacteria (Lactobacillus plantarum ACA-DC 287, L. 146, Lactobacillus paracasei 4037, sakei LQC 1378, and Leuconostoc mesenteroides 1398) was investigated in controlled fermentation cv. Conservolea green olives with a multilayer perceptron network, combined logistic-Fermi function, two-term Gompertz function. Neural network training based on the steepest-descent gradient learning algorithm. Model performance compared experimental data statistical indices, namely coefficient determination (R2), root mean square error (RMSE), relative percentage (MRPE), absolute (MAPE), standard prediction (SEP). set consisted 125 counts (CFU per milliliter) during olive process for up to 38 days (5 × 25 sampling days). For model development, methodology fo...