目前常用Proteinchip Software 3．0检测质谱 片随后在芯片阅读器上进行分析，目前常用且比较成峰，对质谱粗图（raw spectra）进行评估以选取进一步 熟的是飞行时间质谱技术。整个过程一般可在几十分析所需的峰。
基于20个网页-相关网页
The effects of number of principal component, range of wavelength, differentiation and smoothing of raw spectra on model were investigated.
研究主成分数、建模谱区、求导和平滑预处理技术对定量分析模型的影响。
For all samples, the raw spectra were pretreated as second derivative and Norris derivative filter(7,5). partial least square (PLS) was used as the calibration algorithm.
在所有的校正模型中，原始谱图经过二阶导数和偏最小二乘（ PLS）处理，20个外部样品用于所建模型的验证，实际和预测的平均相对偏差<5%。
Spectra of 700 typical samples were collected to develop the calibration models, among which the raw spectra were pretreated with first derivative and partial least square.
收集700个具有代表性样品的光谱，建立其相应指标的近红外模型。
应用推荐