Nowadays, supercritical fluid extraction (SFE) has diverse applications in the areas of natural product extraction and food processing. Thus, many theoretical development, experimental and application studies have been done. Theoretical developments in this field have included the application of statistical mechanical models [1], equation of state methods [2] and a diverse array of solution thermodynamic concepts [3] which explain and correlate the solubility and phase equilibria. These approaches need extensive physicochemical data and are limited to structurally-simple solutes .Currently these theoretical methods have limited use [4]. We have applied structural parameters in uniqueness, binary and triplex combinations to explain solubility in supercritical carbon dioxide. These structural parameters were estimated by Joback, Lydersen, Klincewicz simpled and Klincewicz extended methods [5]. One advantage of this approach is the availability of required data, so one can estimate which parameters affect on maximum solute solubility. This paper proposes a new method, Adaptive Neuro-Fuzzy Inference System (ANFIS) to evaluate structural descriptors of certain organic compounds for their appropriate solubility in terms of QSAR models with the aid of artificial neural network (ANN) approach combined with the principle of fuzzy logic. The ANFIS was utilized to predict solubility which accounts for non-linearities. A data set of 21 compounds was used [6]. The data of 21 organic compounds used to choose the best parameter (s) estimating solubility with less error. Additionally, how the effect of the estimation method accuracy (Joback, Lydersen, Klincewicz simpled and Klincewicz extended methods) on the final results is investigated.

Keywords