Comparison of Machine Learning Methods towards Developing Interpretable Polyamide Property Prediction

Polyamides are often used for their superior thermal, mechanical, and chemical properties. They form a diverse set of materials that have a large variation in properties between linear to aromatic compounds, which renders the traditional quantitative structure–property relationship (QSPR) challengin...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Franklin Langlang Lee, Jaehong Park, Sushmit Goyal, Yousef Qaroush, Shihu Wang, Hong Yoon, Aravind Rammohan, Youngseon Shim
Formato:	article
Lenguaje:	EN
Publicado:	MDPI AG 2021
Materias:	machine learning 1 polyamide 2 QSPR 3 Organic chemistry QD241-441
Acceso en línea:	https://doaj.org/article/ee58647ad139413381d2a2c9efa14436
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:ee58647ad139413381d2a2c9efa14436
record_format	dspace
spelling	oai:doaj.org-article:ee58647ad139413381d2a2c9efa144362021-11-11T18:42:46ZComparison of Machine Learning Methods towards Developing Interpretable Polyamide Property Prediction10.3390/polym132136532073-4360https://doaj.org/article/ee58647ad139413381d2a2c9efa144362021-10-01T00:00:00Zhttps://www.mdpi.com/2073-4360/13/21/3653https://doaj.org/toc/2073-4360Polyamides are often used for their superior thermal, mechanical, and chemical properties. They form a diverse set of materials that have a large variation in properties between linear to aromatic compounds, which renders the traditional quantitative structure–property relationship (QSPR) challenging. We use extended connectivity fingerprints (ECFP) and traditional QSPR fingerprints to develop machine learning models to perform high fidelity prediction of glass transition temperature (<inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><msub><mi>T</mi><mi>g</mi></msub></mrow></semantics></math></inline-formula>), melting temperature (<inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><msub><mi>T</mi><mi>m</mi></msub></mrow></semantics></math></inline-formula>), density (<inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mi>ρ</mi></semantics></math></inline-formula>), and tensile modulus (<inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mi>E</mi></semantics></math></inline-formula>). The non-linear model using random forest is in general found to be more accurate than linear regression; however, using feature selection or regularization, the accuracy of linear models is shown to be improved significantly to become comparable to the more complex nonlinear algorithm. We find that none of the models or fingerprints were able to accurately predict the tensile modulus <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mi>E</mi></semantics></math></inline-formula>, which we hypothesize is due to heterogeneity in data and data sources, as well as inherent challenges in measuring it. Finally, QSPR models revealed that the fraction of rotatable bonds, and the rotational degree of freedom affects polyamide properties most profoundly and can be used for back of the envelope calculations for a quick estimate of the polymer attributes (glass transition temperature, melting temperature, and density). These QSPR models, although having slightly lower prediction accuracy, show the most promise for the polymer chemist seeking to develop an intuition of ways to modify the chemistry to enhance specific attributes.Franklin Langlang LeeJaehong ParkSushmit GoyalYousef QaroushShihu WangHong YoonAravind RammohanYoungseon ShimMDPI AGarticlemachine learning 1polyamide 2QSPR 3Organic chemistryQD241-441ENPolymers, Vol 13, Iss 3653, p 3653 (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	machine learning 1 polyamide 2 QSPR 3 Organic chemistry QD241-441
spellingShingle	machine learning 1 polyamide 2 QSPR 3 Organic chemistry QD241-441 Franklin Langlang Lee Jaehong Park Sushmit Goyal Yousef Qaroush Shihu Wang Hong Yoon Aravind Rammohan Youngseon Shim Comparison of Machine Learning Methods towards Developing Interpretable Polyamide Property Prediction
description	Polyamides are often used for their superior thermal, mechanical, and chemical properties. They form a diverse set of materials that have a large variation in properties between linear to aromatic compounds, which renders the traditional quantitative structure–property relationship (QSPR) challenging. We use extended connectivity fingerprints (ECFP) and traditional QSPR fingerprints to develop machine learning models to perform high fidelity prediction of glass transition temperature (<inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><msub><mi>T</mi><mi>g</mi></msub></mrow></semantics></math></inline-formula>), melting temperature (<inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><msub><mi>T</mi><mi>m</mi></msub></mrow></semantics></math></inline-formula>), density (<inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mi>ρ</mi></semantics></math></inline-formula>), and tensile modulus (<inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mi>E</mi></semantics></math></inline-formula>). The non-linear model using random forest is in general found to be more accurate than linear regression; however, using feature selection or regularization, the accuracy of linear models is shown to be improved significantly to become comparable to the more complex nonlinear algorithm. We find that none of the models or fingerprints were able to accurately predict the tensile modulus <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mi>E</mi></semantics></math></inline-formula>, which we hypothesize is due to heterogeneity in data and data sources, as well as inherent challenges in measuring it. Finally, QSPR models revealed that the fraction of rotatable bonds, and the rotational degree of freedom affects polyamide properties most profoundly and can be used for back of the envelope calculations for a quick estimate of the polymer attributes (glass transition temperature, melting temperature, and density). These QSPR models, although having slightly lower prediction accuracy, show the most promise for the polymer chemist seeking to develop an intuition of ways to modify the chemistry to enhance specific attributes.
format	article
author	Franklin Langlang Lee Jaehong Park Sushmit Goyal Yousef Qaroush Shihu Wang Hong Yoon Aravind Rammohan Youngseon Shim
author_facet	Franklin Langlang Lee Jaehong Park Sushmit Goyal Yousef Qaroush Shihu Wang Hong Yoon Aravind Rammohan Youngseon Shim
author_sort	Franklin Langlang Lee
title	Comparison of Machine Learning Methods towards Developing Interpretable Polyamide Property Prediction
title_short	Comparison of Machine Learning Methods towards Developing Interpretable Polyamide Property Prediction
title_full	Comparison of Machine Learning Methods towards Developing Interpretable Polyamide Property Prediction
title_fullStr	Comparison of Machine Learning Methods towards Developing Interpretable Polyamide Property Prediction
title_full_unstemmed	Comparison of Machine Learning Methods towards Developing Interpretable Polyamide Property Prediction
title_sort	comparison of machine learning methods towards developing interpretable polyamide property prediction
publisher	MDPI AG
publishDate	2021
url	https://doaj.org/article/ee58647ad139413381d2a2c9efa14436
work_keys_str_mv	AT franklinlanglanglee comparisonofmachinelearningmethodstowardsdevelopinginterpretablepolyamidepropertyprediction AT jaehongpark comparisonofmachinelearningmethodstowardsdevelopinginterpretablepolyamidepropertyprediction AT sushmitgoyal comparisonofmachinelearningmethodstowardsdevelopinginterpretablepolyamidepropertyprediction AT yousefqaroush comparisonofmachinelearningmethodstowardsdevelopinginterpretablepolyamidepropertyprediction AT shihuwang comparisonofmachinelearningmethodstowardsdevelopinginterpretablepolyamidepropertyprediction AT hongyoon comparisonofmachinelearningmethodstowardsdevelopinginterpretablepolyamidepropertyprediction AT aravindrammohan comparisonofmachinelearningmethodstowardsdevelopinginterpretablepolyamidepropertyprediction AT youngseonshim comparisonofmachinelearningmethodstowardsdevelopinginterpretablepolyamidepropertyprediction
_version_	1718431797226766336

Comparison of Machine Learning Methods towards Developing Interpretable Polyamide Property Prediction

Ejemplares similares