Open AccessQuantitative Structure-property Relationship Model Based on Artificial Neural Network
Author(s) -
Yue Mu,
Li Sun
Publication year - 2022
Publication title -
journal of engineering research and reports
Language(s) - English
Resource type - Journals
ISSN - 2582-2926
DOI - 10.9734/jerr/2022/v22i517536
Subject(s) - quantitative structure–activity relationship , artificial neural network , robustness (evolution) , artificial intelligence , computer science , machine learning , biological system , biochemical engineering , engineering , chemistry , biochemistry , biology , gene
Artificial neural network (ANN) has been widely researched and applied in chemical process, because of its parallel processing and excellent nonlinear mapping ability, with strong robustness and fault tolerance. By using artificial neural network to establish the model between the properties of mixture and its molecular structure, more accurate data can be predicted and obtained than those determined by experiment. This paper summarizes the development process of artificial neural network and analyzes the application of ANN in quantitative structure-property relationship model (QSPR). It is pointed out that QSPR model combined with artificial neural network can effectively predict the properties of compounds or mixtures, which can shorten the experimental testing process and is able to be widely used with less limitation. It has important significance in application of new biomass fuel, the analysis of the pollution, prediction of the risk of dangerous chemical properties and so on. In the future, there will be broader application space of ANN-QSPR model.