论文部分内容阅读
本文将人工神经网络（ＡＮＮ）用于有机环境污染物紫外光谱库检索。对该神经网络的参数优化作了讨论。并用ＡＮＮ对噪声、杂质等因素的影响作了详细的考察。为了提高紫外光谱的分辨，本文提出用导数光谱作ＡＮＮ训练和检索，使网络的收敛速度明显加快，对检验光谱中杂质的容允程度明显增加。本文还将ＡＮＮ与传统的相关系数法作了比较。结果表明，ＡＮＮ法在抗噪声和杂质等方面明显优于相关系数法。
Artificial neural network (ANN) is used to search the database of organic pollutants. The parameter optimization of this neural network is discussed. And with ANN on noise, impurities and other factors were investigated in detail. In order to improve the resolution of ultraviolet spectrum, this paper proposes to use the derivative spectrum for ANN training and retrieval, the convergence speed of the network is obviously accelerated, and the degree of tolerance to the impurity in the test spectrum is obviously increased. This article also compares ANN with the traditional correlation coefficient method. The results show that the ANN method is superior to the correlation coefficient method in terms of noise immunity and impurities.