Effects of components on and predictive modeling of microemulsion phase behavior in nonionic systems.

This study aimed to investigate phase behaviors, to study effects of cosolvent addition on size of microemulsion regions and to propose modified logistic regression which could describe microemulsion regions in nonionic systems. The systems composed of rice bran oil (RBO) or isopropyl palmitate (IPP), various ratios of sorbitan monooleate (SMO) and polyoxyethylene 20 sorbitan monooleate (PSMO) mixtures, water and isopropyl alcohol (IPA) or propylene glycol (PG) were studied for their microemulsion regions obtained on the phase diagrams. Concept of modified logistic regression was used to predict probability of microemulsion formation and size of microemulsion regions in the systems. It was found that both oil and cosolvent types affected on microemulsion formation. A system composed of IPP, 2:1 water:IPA, and 1:1 SMO:PSMO could provide the largest microemulsion region. However, the purposed modified logistic regression could be used consistently for only one system of the total four systems due to the faceted shape of microemulsion-zone.

Physicochemical transfer functions to predict zone of precipitation between calcium and phosphate in simplified parenteral nutrition.

We use committee-vote via empirical transfer functions obtained from various regression models of actual physicochemical data for describing possible zone of hazardous precipitation of calcium and phosphate for particular amino acid brand name, i.e., Moripron-F® or Aminovenos-N-Paed®. The logistic regression with modified Hosmer-Lemeshow method is also presented. We use a slack variable for grouping data space near each original data point for calculating odd ratio of precipitation event and then extrapolate back to zero value of the slack variable in order to obtain the intrinsic odd ratio function that is independent from the slack variable. Committee-vote could reveal a nonlinear pattern with high accuracy and low health hazardous prediction for Moripron-F®. However, applying such algorithm with Aminovenos-NPaed ® was unsuccessful due to violation of the committee-voting rule, i.e., empirical transfer function has too weak prediction power.

Using Algebraic Geometry to Describe Boundaries of Microemulsion Regions in Nonionic Systems.

This study aimed to propose an empirical polynomial function based on algebraic geometry which could represent microemulsion regions in nonionic systems. The systems composed of olive oil (OO), water (W) and various ratios of sorbitan monooleate (SMO) and polyoxyethylene 20 sorbitan monooleate (PSMO) mixtures were investigated for their microemulsion regions obtained on the phase diagrams. Concept of algebraic geometry in this study was that description of close-boundary shape was cross-sectioning of a 3D-geometrical object with a plane which could be explained by an empirical polynomial function of the {oil, water, surfactant mixture}. In-house ad hoc software was created according to the proposed concept. It was found that this method could describe boundaries of microemulsion regions in the studied systems. Although this method could not reduce time consumption since it required intensive computational powers, the software could automatically run. Moreover, the developed program could provide acceptable level of accuracy.