|
|
Detection of pH variable in solidstate fermentation process by FTNIR spectroscopy and BPAdaboost |
School of Electrical and Information Engineering, Jiangsu University, Zhenjiang, Jiangsu 212013, China |
|
|
Abstract In order to achieve rapid detection of pH value in solidstate fermentation process, a new method was proposed based on nearinfrared spectroscopy. The nearinfrared spectroscopy of 140 solidstate fermented product samples were obtained within the wavelength range of 10 000 to 4 000 cm-1, and the reference measurement of pH were achieved by pH meter. Combined with backpropagation (BP) neural network, the adaptive boosting algorithm (Adaboost) was employed to establish BPAdaboost model with 10 weak predictors. The experimental results show that the proposed model has good prediction accuracy and correlation with root mean square error RMSEP of 0.072 6 and correlation coefficient R of 0.981 1, respectively.
|
Received: 27 June 2012
|
|
|
|
|
|
|