Small Data Machine Learning: Classification and Prediction of Poly(ethylene terephthalate) Stabilizers Using Molecular Descriptors

Abstract

Experimental data from a patent were analyzed to learn about the small molecule additives that were most effective in mitigating the degradation of polyethylene terephthalate. Two sets of molecular descriptors were calculated for a dataset of 39 additive candidates; unsupervised and supervised analyses were performed to determine the most influential structural features that led to reduced degradation. A clustering approach revealed evidence that performance differences had some structural pattern dependence on the molecular descriptors that were employed. To pinpoint the features responsible for those physical differences, a reduced design region approach was applied to analyze descriptors both individually and in multiple dimensions to determine the effectiveness in a binary classification of high and low performances. For each molecular descriptor type, two or three influential descriptors were identified and justified with respect to the additive performance and physicochemical ability to mitigate degradation. Random forest models were constructed with relatively high predictability for both MACCS-166 (AUC = 0.86) and alvaDesc molecular descriptors (AUC = 0.93). We compare molecular descriptor methods for their ability to construct classifiers and to prioritize experimental work toward building a rich dataset. We find that, in small materials datasets, understanding the underlying physicochemical behavior is indispensable for validating the effectiveness of machine learning models.

Data Science Guided Experiments Identify Conjugated Polymer Solution Concentration as a Key Parameter in Device Performance

“…The list of publications utilized for curating this data set is provided in the Supporting Information. The classification algorithm developed by McBride et al , was applied to the curated data set to identify the most impactful processing variables from a list of 21 numerical variables that confine the devices with a “high” field effect transistor (FET) hole mobility (>1 cm 2 /V·s) to a reduced design space. The advantage of this algorithm is that it selects the key processing parameters from a long list of variablesthus preventing the need to investigate all the processing variables individually, which can be tedious and time-consuming.…”

System-Level Network Analysis of a Catechol Component for Redox Bioelectronics

“…The results in Figure a indicate that four of these metrics are expected to be sensitive to topology (ΔCharge, ΔA 480 or Δ n Cat ox , Q ox Fc , and Q red Ru ), while one metric has a weaker topology dependence (ΔCurrent). While these four sensitive metrics may be useful individually, combinations of metrics often offer greater discriminating abilities through multivariant and data-driven machine learning analysis. ,,− …”

Application of Machine Learning in Developing Quantitative Structure–Property Relationship for Electronic Properties of Polyaromatic Compounds

“…These descriptors can be as simple as a list of molecule’s compositions 5 , 6 or as complex as matrices, 18 − 20 fingerprints, 15 , 16 , 21 etc. 8 , 14 , 22 − 26 …”

“…Typically, these descriptors are numerical objects obtained by transforming raw molecular data by some predefined procedure. These descriptors can be as simple as a list of molecule’s compositions , or as complex as matrices, − fingerprints, ,, etc. ,,− …”