| Title |
Prediction of protein and lipid content in black soldier fly (Hermetia illucens L.) larvae flour using portable NIR spectrometers and chemometrics |
| ID_Doc |
8437 |
| Authors |
Cruz-Tirado, JP; Vieira, MSD; Amigo, JM; Siche, R; Barbin, DF |
| Title |
Prediction of protein and lipid content in black soldier fly (Hermetia illucens L.) larvae flour using portable NIR spectrometers and chemometrics |
| Year |
2023 |
| Published |
|
| DOI |
10.1016/j.foodcont.2023.109969 |
| Abstract |
Black soldier fly (BSF) larvae meet circular economy requirements by transforming waste into high-quality protein and lipids. Because of the rapid larva development to the mature stage (14 days - 2 months), the insect production industry seeks rapid analytical methods. We evaluated the performance of two portable NIR spectrometers (1: 900-1700 nm; 2: 1350-2562 nm), coupled to Partial Least Square Regression (PLSR) and Support Vector Machine Regression (SVMR) to predict protein and lipid content (%) in BSF larvae flour. The spectra dataset was explored by Principal Component Analysis (PCA). PLSR and SVMR performed similarly in predicting protein content for both spectrometers according to Residual Prediction Deviation (RPD >2.5) and Root Mean Square Error of Prediction (RMSEP = 1.9%). SVMR turned out to yield a better prediction performance for the lipidic content (RMSEP = 3.51%; RPD = 4.32) respect to PLSR. Moreover, spectrometer 2, working at a higher wavelength range, showed better performance than spectrometer 1. In addition, a variable selection step was performed, where interval PLS (iPLS) and genetic algorithm (GA) improved PLSR models. In conclusion, a portable NIR spectrometer coupled with chemometrics could support rapid analytical measurements in the insect industry. |
| Author Keywords |
Insect; Near infrared; Machine learning; Variable selection |
| Index Keywords |
Index Keywords |
| Document Type |
Other |
| Open Access |
Open Access |
| Source |
Science Citation Index Expanded (SCI-EXPANDED) |
| EID |
WOS:001047029500001 |
| WoS Category |
Food Science & Technology |
| Research Area |
Food Science & Technology |
| PDF |
|