A Computer Aided Framework for Prediction of Properties of Organic Systems

This paper presents a systematic computer aided framework for modeling and prediction of pure component and mixture properties for a very wide range of chemicals and properties. Together with the extended GC property models for pure component properties and mixtures, the framework provides a feature for prediction of missing group interaction parameters (for mixture properties) by employing an atom-connectivity approach. The framework also provides a model parameter regression feature to fine-tune existing parameters and regress new parameters.