Résumé
<p><b>OBJECTIVE</b>To investigate into the method of "multi activity index evaluation and combination optimized of mult-component" for Chinese herbal formulas.</p><p><b>METHOD</b>According to the scheme of uniform experimental design, efficacy experiment, multi index evaluation, least absolute shrinkage, selection operator (LASSO) modeling, evolutionary optimization algorithm, validation experiment, we optimized the combination of Jiangzhi granules based on the activity indexes of blood serum ALT, ALT, AST, TG, TC, HDL, LDL and TG level of liver tissues, ratio of liver tissue to body.</p><p><b>RESULT</b>Analytic hierarchy process (AHP) combining with criteria importance through intercriteria correlation (CRITIC) for multi activity index evaluation was more reasonable and objective, it reflected the information of activity index's order and objective sample data. LASSO algorithm modeling could accurately reflect the relationship between different combination of Jiangzhi granule and the activity comprehensive indexes. The optimized combination of Jiangzhi granule showed better values of the activity comprehensive indexed than the original formula after the validation experiment.</p><p><b>CONCLUSION</b>AHP combining with CRITIC can be used for multi activity index evaluation and LASSO algorithm, it is suitable for combination optimized of Chinese herbal formulas.</p>