Predicting the heat capacities of ammonium- and phosphonium-based deep eutectic solvents using artificial neural network

Abstract Deep eutectic solvents (DESs) are more environment-friendly and sustainable than ionic liquids (ILs). Determining their physical properties is vital prior to application. However, it not practical experimentally determine the of all DESs. Therefore, developing models estimate necessary. In this study, a generalized model using artificial neural network (ANN) was developed predict heat capacities ammonium- phosphonium-based DESs over temperature range 278.15 – 353.15 K at atmospheric pressure. The best ANN from training optimization process has an architecture with two hidden layers 7 6 neurons, respectively. overall average absolute relative deviation proposed data 0.57%. Qualitative analyses suggest promising predicting capability model. calculated values statistical descriptors close ideal values. can be used reliably other two-component