Prediction of properties of mixture by integration of pattern recognition and logic procedure by using fuzzy clustering

摘要

New methodology is presented here to overcome the difficult in predicting non-linear cooperative effects caused by mixing components. One example in which such difficulties are found is the prediction of properties of mixed solvents. This work integrates two methodologies, fuzzy clustering as pattern recognition and logical procedure. Virtual meta-probability space and Lebesgue integral are adopted to develop mathematical formulation. This work provides an answer for the calculation of properties of mixed components, such as solubility from the data of prototypes in binary system. Development of applications to other objects is expected.