光谱学与光谱分析
光譜學與光譜分析
광보학여광보분석
SPECTROSCOPY AND SPECTRAL ANALYSIS
2014年
7期
1844-1848
,共5页
朱哲燕%刘飞%张初%孔汶汶%何勇
硃哲燕%劉飛%張初%孔汶汶%何勇
주철연%류비%장초%공문문%하용
中红外光谱%香菇%蛋白质含量%连续投影算法
中紅外光譜%香菇%蛋白質含量%連續投影算法
중홍외광보%향고%단백질함량%련속투영산법
Mid-infrared spectroscopy%Shiitake mushroom%Protein content%Successive projections algorithm
研究了中红外光谱预测香菇蛋白质含量的可行性。去掉明显噪声部分后，研究香菇3581～689 cm -1中红外光谱与蛋白质含量的关系。以Savitzky-Golay （SG ）5点平滑预处理光谱建立偏最小二乘法（par-tial least squares ，PLS）的预测模型的效果不理想，模型的建模集和预测集的相关系数均高于0.85，但剩余预测偏差（residual prediction deviation ，RPD）值仅为1.77。采用连续投影算法（successive projections algo-rithm ，SPA ）算法从3000个波数点中选择7个特征波数，并以七个特征波数分别建立PLS、多元线性回归（multiple linear regression ，MLR）、反向传播神经网络（back-propagation neural network ，BPNN）和极限学习机模型（extreme learning machine ，ELM）。与全谱的PLS相比，以特征波数的PLS模型和MLR模型的预测效果相对较差，而以特征波数的BPNN和ELM模型的预测效果相对较好。其中SPA-ELM模型的预测效果最佳，预测集相关系数（correlation coefficient of prediction ）Rp ＝0.8995，预测集均方根误差（root mean square error of prediction）RMSEP＝1.4313，剩余预测偏差RPD＝2.18。研究结果表明，中红外光谱分析技术可以用于预测香菇蛋白质含量，且SPA选取特征波数能用来代替原始光谱进行建模分析，为香菇蛋白质含量的检测提供了新的思路。
研究瞭中紅外光譜預測香菇蛋白質含量的可行性。去掉明顯譟聲部分後，研究香菇3581～689 cm -1中紅外光譜與蛋白質含量的關繫。以Savitzky-Golay （SG ）5點平滑預處理光譜建立偏最小二乘法（par-tial least squares ，PLS）的預測模型的效果不理想，模型的建模集和預測集的相關繫數均高于0.85，但剩餘預測偏差（residual prediction deviation ，RPD）值僅為1.77。採用連續投影算法（successive projections algo-rithm ，SPA ）算法從3000箇波數點中選擇7箇特徵波數，併以七箇特徵波數分彆建立PLS、多元線性迴歸（multiple linear regression ，MLR）、反嚮傳播神經網絡（back-propagation neural network ，BPNN）和極限學習機模型（extreme learning machine ，ELM）。與全譜的PLS相比，以特徵波數的PLS模型和MLR模型的預測效果相對較差，而以特徵波數的BPNN和ELM模型的預測效果相對較好。其中SPA-ELM模型的預測效果最佳，預測集相關繫數（correlation coefficient of prediction ）Rp ＝0.8995，預測集均方根誤差（root mean square error of prediction）RMSEP＝1.4313，剩餘預測偏差RPD＝2.18。研究結果錶明，中紅外光譜分析技術可以用于預測香菇蛋白質含量，且SPA選取特徵波數能用來代替原始光譜進行建模分析，為香菇蛋白質含量的檢測提供瞭新的思路。
연구료중홍외광보예측향고단백질함량적가행성。거도명현조성부분후，연구향고3581～689 cm -1중홍외광보여단백질함량적관계。이Savitzky-Golay （SG ）5점평활예처리광보건립편최소이승법（par-tial least squares ，PLS）적예측모형적효과불이상，모형적건모집화예측집적상관계수균고우0.85，단잉여예측편차（residual prediction deviation ，RPD）치부위1.77。채용련속투영산법（successive projections algo-rithm ，SPA ）산법종3000개파수점중선택7개특정파수，병이칠개특정파수분별건립PLS、다원선성회귀（multiple linear regression ，MLR）、반향전파신경망락（back-propagation neural network ，BPNN）화겁한학습궤모형（extreme learning machine ，ELM）。여전보적PLS상비，이특정파수적PLS모형화MLR모형적예측효과상대교차，이이특정파수적BPNN화ELM모형적예측효과상대교호。기중SPA-ELM모형적예측효과최가，예측집상관계수（correlation coefficient of prediction ）Rp ＝0.8995，예측집균방근오차（root mean square error of prediction）RMSEP＝1.4313，잉여예측편차RPD＝2.18。연구결과표명，중홍외광보분석기술가이용우예측향고단백질함량，차SPA선취특정파수능용래대체원시광보진행건모분석，위향고단백질함량적검측제공료신적사로。
The feasibility of protein determination of shiitake mushroom (Lentinus edodes ) using mid-infrared spectroscopy (MIR) was studied in the present paper. Wavenumbers 3 581~689 cm -1 were used for quantitative analysis of protein content after removing of the part of obvious noises. Five points Savitzky-Golay smoothing was applied to pretreat the MIR spectra and partial least squares (PLS) model was built based on the pretreated spectra. The full spectra PLS model obtained poor perform-ance with the ratio ofprediction to deviation (RPD) of only 1.77. Successive projections algorithm (SPA) was applied to select 7 sensitive wavenumbers from the full spectra ,and PLS model ,multiple linear regression (MLR) ,back-propagation neural net-work (BPNN) and extreme learning machine (ELM) model were built using the selected sensitive wavenumbers. SPA-PLS mod-el and SPA-MLR model obtained relatively worse results than SPA-BPNN model and SPA-ELM model. SPA-ELM obtained the best results with correlation coefficient of prediction (Rp ) of 0.899 5 ,root mean square error of prediction (RMSEP) of 1.431 3 and RPD of 2.18. The overall results indicated that MIR combined with chemometrics could be used for protein content determi-nation of shiitake mushroom ,and SPA could select sensitive wavenumbers to build more accurate models instead of the full spec-tra.
