APPLICATION OF FAM NETWORKS IN RECOMMENDATION OF MOBILE PHASE FOR RP-HPLC

AIM To establish expert system to recommend solvent strength in RP-HPLC by using FAM (fuzzy associate memorizer) networks. METHODS Symbolic rules reflecting the quantitative relationship among molecular structure, capacity factors and solvent strength in RP-HPLC were extracted from trained radial basis function networks. Then FAM networks consisting of these rules were built to recommend mobile phase strength for RP-HPLC. RESULTS The validation results of the system by monocomponent and mutilcomponent drug samples were satisfied. CONCLUSION The mobile phase recommendation system for RP-HPLC analysis of drugs showed reliable performance.