Comparison of Machine Learning Methods towards Developing Interpretable Polyamide Property Prediction

Abstract

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 (Tg), melting temperature (Tm), density (ρ), and tensile modulus (E). 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 E, 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.

Machine-Learning-Based Predictions of Polymer and Postconsumer Recycled Polymer Properties: A Comprehensive Review

“…Lee et al 150 collected the data on polyamide structures from the PoLyInfo database and predicted the properties such as density and melting and glass transition temperatures using a multiple linear regression algorithm. To carry out high-fidelity polyamide property predictions, the authors used traditional QSPR fingerprints and extended connectivity fingerprints to develop ML models.…”

Atom-Based Machine Learning Model for Quantitative Property–Structure Relationship of Electronic Properties of Fusenes and Substituted Fusenes

“…Modern cheminformatics provides several widely used, general-purpose 2D molecular fingerprints that summarize molecular structures as collections of substructures. These fingerprints include the Molecular ACcess System (MACCS) , and the extended connectivity fingerprint (ECFP) algorithm for structural representation in various modeling disciplines. ,,,− However, these fingerprints are binary vectors that indicate the presence or absence of certain structural features to fully capture the structures of molecules that contain repetitive or similar subunits, such as polymers . This is a particularly relevant concern for polycyclic aromatic hydrocarbons (PAHs) which are constructed by attaching a number of benzene units, and their electronic properties are strongly influenced by their size and shape.…”

New Opportunity: Machine Learning for Polymer Materials Design and Discovery

“…Making full use of these data by ML could figure out some patterns hidden in data and construct quantitative structure–property relationship (QSPR) models for predicting the properties of unknown materials to design new materials with desired target properties. [ 9–14 ]…”

“…Making full use of these data by ML could figure out some patterns hidden in data and construct quantitative structure-property relationship (QSPR) models for predicting the properties of unknown materials to design new materials with desired target properties. [9][10][11][12][13][14] As one of the most important branches in materials, polymers have kept playing a key part for the various macromolecular structures and architectural properties. [15][16][17] Polymers have been widely used in both consumer products and engineering applications such as aviation, automobiles, ships, infrastructure, military supplies, and many other fields.…”