Abstract
Quantitative structure property relationships (QSPR) mine data sets of experimentally measured properties to predict different material properties through mathematical regression and machine learning. QSPR have the capability to predict any property, as long as there is a large enough data set of measured values. It is often costly and time-consuming to experimentally test certain properties of ionic liquids, and modern fundamental theoretical (quantum chemical or molecular dynamics) approaches require significant computer resources and may have limited accuracy. On the other hand, QSPR allow for simulations to converge to a result in just a few hours. These simulations are reliable, fast, and accurate, and only require a small number of input parameters. In this work, we predict surface tension and electrical conductivity values of potential ionic liquid propellants via QSPR, and present a detailed discussion on the error analysis performed on the results obtained.