Research On Expert System For Properties Regulation Of Traditional Chinese Medicine Tablets Applied To Direct Compression

Objective:This study aimed to（ⅰ）Analyze the key parameters related to tablet quality by determining the basic physical and functional properties of Traditional Chinese Medicine（TCM）powder（extracts,direct pulverization）and excipients;（ⅱ）Analyze the deformation mechanism and compaction mechanism of powder during compression;（ⅲ）Study the regulations of tablet quality;（ⅳ）Preliminarily establish an expert system with functions such as evaluation of powder direct compression（DC）,recommendation of excipients and prescriptions,and regulate of tablet quality.This expert system can provide convenience and basis for the research of DC preparation.Methods:（1）44 types of TCM powders and 18 types of DC excipients were selected for research.The basic physical properties,functional properties,tableting phenomenon,and tablet quality of the powder were characterized and analyzed.The relationship between the basic physical properties of the powder and the quality of the tablet is analyzed by statistical methods.（2）Based on functional properties,principal component analysis（PCA）and hierarchical cluster analysis（HCA）are used to classify the powders and analyzed the characteristics of each category of powder.And the powder classification rules are constructed through the decision tree J48 algorithm.In addition,partial least squares（PLS）regression models are established to predict the tablet quality based on the powder’s functional property.（3）Using the mixing design,single factor investigation,and other experimental designs combined with machine learning algorithms to analyze and explore the types and dosages of commonly used tablet excipients（fillers,disintegrants,lubricants,and glidants）on the regulation of tablet quality,and establish formula design rules for powder DC.（4）Based on the above research results and the basic thinking and framework of the formulation research,the overall reasoning framework of the formulation design system is constructed.Java is used to program the rules and reasoning framework obtained in the preliminary research,and to initially construct the recommendation expert system.Five different types of TCM powder were input into the system for operation,and verify the operating results through experiments.Results:（1）The measurement method of basic physical properties（moisture content,true density,particle size,particle size distribution,etc.）is determined.The physical properties of TCM extract powder are slightly different from direct pulverization and fillers,while the TCM extracts are similar.The qualities of tablets prepared with TCM powder have certain differences.Among them,78%of the extract was sticking,59.1%of the powder is unqualified for weight variation,and 4.5%of the powder is unqualified disintegrated.However,the correlation between the basic physical properties of traditional Chinese medicine powder and tablet quality is weak.In addition,the characterization and calculation methods of functional properties（compressibility,compactibility,elasticity）were determined,and a new model(CR=y₀+Ae^{-b P})was established to simultaneously characterize the compressibility and compactibility of the powder.（2）Nine functional parameters(compressibility（w,a,k _G,y₀）,compactibility（k_a,k_b,b）,elasticity(k _FES,k _E3))are determined.The powders could be classified well based on their functional parameters.The order of powders category suitable for DC isⅠ>Ⅱ>Ⅲ>Ⅳ>Ⅴ.Powders inⅠcategory might be eligible for DC requirements,whileⅤcategory should be modified or granulated to meet DC.Ⅱcategory,Ⅲcategory,andⅣcategory would be adequate to DC through reasonable formulation design.Parameters w,k_G,b,k_a are the critical quality attributes.w,k _G,b,and TS,k_a and DT,and w,k _G,k_a,and SF have a significant correlation.（3）The applicability rules and quality control rules for DC of TCM have been constructed,including filler screening rules,disintegrant screening rules,tablet weight variation regulation rules,prediction sticking rules,lubricant screening rules.These rules basically meet the needs of tablet quality regulation.（4）Established a fair framework to effectively integrate the above rules.An Expert System（http://106.15.1.18/tablet/login）for quality regulation of TCM tablets apply to DC has been preliminarily constructed and all parts are operating normally.The verified experiments were conducted.It is resulted that the tablets prepared by using the formulation recommended by the system all meet the standard.Conclusion:In this study,based on the functional properties of traditional Chinese medicine powder,through data analysis and mining,the correlation between the functional properties of the powder and the quality of the tablet was explored through date analysis and mining,and the critical quality attributes were determined.DC applicability rules and tablet qualities regulation rules are constructed by machine learning algorithms.On this basis,the expert system is constructed using Java.And the expert design was verified by experiments.The results show that the expert system has basically realized the applicability evaluation,formulation design,and quality regulation functions and goals of DC of TCM powder.