Prediction of the postharvest quality of Boletus wild mushrooms stored with mesoporous silica nanoparticles antibacterial film using Long Short-Term Memory model combined with the Northern Goshawk Optimization (NGO-LSTM)

Yudi Liu,Kai Jiang,Yuyue Qin,Margaret Brennan,Charles Brennan,Jianxin Cao,Zhengxuan Wang,Thanapop Soteyome

DOI: https://doi.org/10.1016/j.foodchem.2024.141490

2024-09-30

Abstract:This study aimed to address the challenge of extending the shelf life of Boletus wild mushrooms, which are prone to environmental and microbial contamination. An antibacterial film composed of polylactic acid (PLA) and mesoporous silica nanoparticles loaded with citral (CMP film) was developed for this purpose. Fifteen quality indices were assessed, and the data were integrated using AHP and TOPSIS to evaluate the film's efficacy. The CMP film effectively maintained the quality of mushroom over time. Additionally, a Nonlinear Global Optimization-Long Short-Term Memory (NGO-LSTM) model was employed to predict storage quality, using seven highly correlated quality indicators. The model achieved a high predictive accuracy, with the R2 exceeding 0.999. This study presents a novel packaging solution and a predictive model that together enhance the storage and quality control of Boletus wild mushrooms.