Escherichia coliBL21 |
Comparing RSM and ANN for optimizing
lipase production |
Glucose, sodium chloride (NaCl), temperature and
induction time |
ANN showed better R2 and
adjusted-R2 values, and lower AAD and RMSE values |
[99] |
Sludge |
Integrating RSM and ANN for improving biohydrogen titer |
Carbon sources, metal cofactor Fe0, pH, and sludge
concentration |
Developing a robust, cost-effective, and reliable
optimization methodology |
[100] |
Kluyveromyces lactis |
Using ANN coupled with GA for media
optimization in hIFN‐γ production |
Medium components (sorbitol,
glycine, Na2HPO4, and MgSO4.7H2O) |
Achieving maximum productivity in
shake flasks level and bioreactor level |
[101] |
Pichia pastoris |
MLP3 neural network was used to optimize the
controller for a novel fed-batch production of A1AT |
Controller gain
and control poles |
The cell growth and protein titer were improved
significantly in comparison to traditional approaches |
[102] |
Rhodotorula glutinis
|
Establishing ANN and SVM models for microbial lipid production
|
Biomass, lipid yield, and COD removal rate
|
SVM performed better than ANN for small samples
Fermentation parameters were optimized by integration of SVM and
GA
|
[103]
|
Sludge |
Developing a hybrid model for optimizing CBP production using
LS-SVM and OED |
Corn stalk weight, ultrasonic duration time,
pretreatment time, and dual-frequency ultrasound |
The model increases
CBP 14.13% more than pure OED |
[104] |
Escherichia coli |
Using random forest and ANN for biomass and
recombinant protein modeling in a fed-batch process |
Temperature,
induction strength, growth rate, soluble/insoluble product formation |
The accuracy reached about ±4% for dry cell mass and ±12% for protein
concentration |
[106] |
saccharomyces cerevisiae |
Employing GPR for prediction of
S. cerevisiae biomass concentration |
Molasses feed rate |
Experimentally validation results showed high accuracy |
[107] |
Penicillium brevicompactum |
Establishing a
k-nearest-neighbor model to optimize different independent
factors in MPA production |
Ultrasound power, irradiation duration,
treatment frequency and duty cycle |
1.64-fold improvement observed in
MPA production |
[108] |