A Priori Knowledge-Incorporating Method for the Determination of Polycyclic Aromatic Hydrocarbons (PAHs) in Edible Vegetable Oils by Time Resolved Fluorescence. (24th May 2022)

• Record Type: Journal Article
• Authors: Chen, Hui; Feng, Meiqin; Li, Jing; Lu, Jie; Gu, Haiyang; Chen, Junhong; He, Shihang; Qi, Xingpu; Chen, Wenjun; Chen, Tong
• Is Part Of: Analytical letters. Volume 55:Number 8(2022)
• Page Start: 1217
• Page End: 1234
• Publication Date: 2022-05-24
• Subjects: Chemometrics -- deep autoencoder -- polycyclic aromatic hydrocarbons -- priori model-incorporating deep-autoencoder models -- time resolved fluorescence; Chemistry, Analytic -- Periodicals; Chemistry, Analytic -- Abstracts; 543
• DOI: 10.1080/00032719.2021.1992417
• Languages: English
• ISSNs: 0003-2719

Abstract

Abstract: A priori knowledge-incorporating method based on time resolved fluorescence was successfully developed for the determination of polycyclic aromatic hydrocarbons in edible vegetable oils. Specifically, fluorescence decay functions of polycyclic aromatic hydrocarbons at characteristic emission wavelengths were used as the priori models and incorporated into the deep-autoencoder. The priori model-incorporating deep-autoencoder models were shown to be effective for the determination of polycyclic aromatic hydrocarbons in edible vegetable oils and root-mean-square errors of prediction lower than 2% were achieved. The influence of analyte, matrix and proportion of priori model were characterized. Increasing the proportion of priori model appropriately was beneficial to the performance of models and 16% was shown to be the best incorporated proportion.