APPLICATION OF MACHINE LEARNING METHODS TO PREDICT THE FLASH POINT OF ORGANIC COMPOUNDS
Authors:
1.
Kazan National Research Technological University
2.
Kazan National Research Technological University
Russian Federation
Russian Federation
3.
Kazan National Research Technological University
Type:
Article
Pages:
from 112
to 117
Status:
In work
Language:
Russian
Keywords:
BIG DATA, INDUSTRY 4.0, FLASHPOINT TEMPERATURE, MACHINE LEARNING, ARTIFICIAL INTELLIGENCE
Abstract (English):
The increase in safety requirements for industrial entities has led to a number of technological changes in the direction of strengthening industrial safety measures. Currently, it is mandatory to develop a safety data sheet for production facilities, especially if chemical production is associated with the use of flammable liquids. At the same time, the use of digital technologies is a priority in creating safe chemical technology systems. This area is part of Industry 4.0. In many countries, the flash point is a measure of the separation of liquids into flammable and flammable. Flash point data for organic compounds were taken from the PubChem database. Information on the flash point for 1,741 organic substances was included in the database for this work. To simplify the analysis of the representation of organic compounds, we used MACC keys, as they are among the best descriptors for predicting the properties of chemical compounds. These descriptors are created based on the shared keys of the substructure. In addition, the models were calculated using Morgan molecular prints, also known as circular prints with a radius of 2. Within the framework of this work, ridge regression, the random forest algorithm, the kNN nearest neighbor method, the support vector machine (SVM) method, and artificial neural networks were implemented. For the training sample, the resulting classification model of a random forest showed an error-free classification, the prediction error for it is 0. The statistical characteristics of the constructed RF model for the sample have the following values: accuracy ACC = 0.83, sensitivity SPC = 0.81, specificity SEN = 0.86, balanced accuracy BA = 0.83, Matthews correlation coefficient MCC = 0.72. On the basis of the developed model, a forecast was made for the belonging of compounds to a certain class ("1" class combustible or "0" flammable) for organic compounds for which there is no experimental information regarding belonging to class "1" or "0".
The increase in safety requirements for industrial entities has led to a number of technological changes in the direction of strengthening industrial safety measures. Currently, it is mandatory to develop a safety data sheet for production facilities, especially if chemical production is associated with the use of flammable liquids. At the same time, the use of digital technologies is a priority in creating safe chemical technology systems. This area is part of Industry 4.0. In many countries, the flash point is a measure of the separation of liquids into flammable and flammable. Flash point data for organic compounds were taken from the PubChem database. Information on the flash point for 1,741 organic substances was included in the database for this work. To simplify the analysis of the representation of organic compounds, we used MACC keys, as they are among the best descriptors for predicting the properties of chemical compounds. These descriptors are created based on the shared keys of the substructure. In addition, the models were calculated using Morgan molecular prints, also known as circular prints with a radius of 2. Within the framework of this work, ridge regression, the random forest algorithm, the kNN nearest neighbor method, the support vector machine (SVM) method, and artificial neural networks were implemented. For the training sample, the resulting classification model of a random forest showed an error-free classification, the prediction error for it is 0. The statistical characteristics of the constructed RF model for the sample have the following values: accuracy ACC = 0.83, sensitivity SPC = 0.81, specificity SEN = 0.86, balanced accuracy BA = 0.83, Matthews correlation coefficient MCC = 0.72. On the basis of the developed model, a forecast was made for the belonging of compounds to a certain class ("1" class combustible or "0" flammable) for organic compounds for which there is no experimental information regarding belonging to class "1" or "0".
Keywords:
BIG DATA, INDUSTRY 4.0, FLASHPOINT TEMPERATURE, MACHINE LEARNING, ARTIFICIAL INTELLIGENCE
BIG DATA, INDUSTRY 4.0, FLASHPOINT TEMPERATURE, MACHINE LEARNING, ARTIFICIAL INTELLIGENCE
