Based Mixture Regression Design Method To Predict The Concentration Ratio Of Blank Matrix

In order to improve the patch's performance,the paper restudies the base material's component contents of the Chinese traditional medicinal patch "Xingshen patch" that is researched by heilongjiang university of chinese medicine.Because of the linear restriction among the base material's components and the restricted condition,this paper designs a uniform program with restricted mixtures that based on the conditional distribution method.It can generate a set of points that can be consonant with the mixtures condition and be more uniform on the experimental domain in order that the less experimental points can represent the more experimental informations.According to every target's feature,I establish the multivariable to multivariable models that been based on partial least-squares method,and the paper adopts least-squares idea to find the multi-target decision-making model of compositive targets, then it takes advantage of sequential linear programming method to obtain the optimal solution,namely:stir temperature is 50 centigrade degree,stir time is 12 hours,the content of base material'shumectant is 45%,the content of moisture is 15%,the content of tackifier is 30%,and the content of excipient is 10%.According to the experimental verification,the base material's compositive performance under the condition is better than the past experimental solution.The patch possesses higher applied value,and it is easier to produce the patch industrializly, so the marketable prospect is broader.Furthermore,the paper can provide with the scientific foundation and theoretical base for homogeneous products.