Deep Learning and Mass Spectrum Based Analysis of Vocs Components

The composition and concentration of Volatile Organic Compounds (VOCs) in the atmosphere can reflect the quality of the air, and the environmental quality changes with the quantity of these compounds. When unknown VOCs are encountered, researchers usually use gas chromatography-mass spectrometry (GC-MS) to measure and analyze them. This discriminative mode requires data analysts with a certain theoretical and practical foundation, is demanding and labor-intensive, and may also introduce errors due to the numerous steps. In order to solve these problems, we propose a deep learning and mass spectrum based method for the analysis of Vocs components. Using the deep learning technique, first, a high-quality mass spectral library is constructed as a reference library using molecular fingerprint information, and then, the sample data obtained in the GC-MS gas chromatography-mass spectrometer is preprocessed with data to extract mass spectra that can represent the VOCs components; finally, the selected candidate mass spectra are library matched with the reference library to return high matching VOCs components results. The experimental results show that the method can accurately and quickly discriminate the components of VOCs.