基于主成分分析的有机化合物吸收光谱识别方法

Identification of organic compounds by absorption spectra based on principal component analysis

某些有机化合物的光谱吸收峰较为隐蔽,很难识别及分析,为此,提出基于主成分分析的有机化合物吸收光谱识别方法。计算有机化合物分子吸收的能量,构建吸收光谱,对获取到的数据进行预处理,利用主成分分析法计算吸收系数谱线中数据点特征,降低数据维度;利用支持向量机分类特征数据,根据输出结果判断其是否属于已知数据库,完成识别过程。选取30种有机化合物作为样本进行仿真实验,结果表明,所提方法具有较好的识别精度,即使是吸收峰极不明显的化合物,依然可以实现有效识别。

The spectral absorption peaks of some organic compounds are obscure and difficult to be identified and analyzed.Therefore,a method of absorption spectrum recognition of organic compounds based on principal component analysis is proposed.The energy absorbed by organic compound molecules was calculated,the absorption spectrum was constructed,and the obtained data were preprocessed.The characteristics of data points in the absorption coefficient spectrum line were calculated by using principal component analysis method,and the first-order,second-order and third-order difference data were obtained to reduce the data dimension.Support vector machine is used to classify feature data and judge whether it belongs to the known database according to the output result.If it belongs to the corresponding category,the recognition process is completed.In the simulation experiment,30 kinds of organic compounds were selected as samples to carry out the simulation experiment,and the results show that the proposed method has ideal recognition accuracy,even the compounds with very little absorption peak can still be recognized effectively.