基于多指标和多元统计分析的百两金综合质量评价OA北大核心CSTPCD

Based on seven indicators(moisture,total ash,acid-insoluble ash,extract,total flavonoids,total phenols and total saponins content),correlation analysis,principal component analysis,clustering heatmap analysis,CRITIC-TOPSIS and CRITIC-grey relational analysis models were used to comprehensively evaluate the quality of seven batches of Ardisia crispa(Thunb.)A.DC.from different areas.The results of correlation analysis showed close relationship among seven indexes,indicating some overlap of information.Principal component analysis extracted five main indicators,including total ash,acid-insoluble ash,extract,total saponins and moisture content,and the cumulative contribution rate was 93.293% .The weighting results of each indicator based on CRITIC method were basically consistent with principal component analysis,among which the weight values of the extract and total saponins content ranked highest.Both evaluation models showed significant quality differences among the A.crispa samples from different areas,but the coefficient of variation of the CRITIC-grey relational analysis model was larger to 42.889 9%,which better reflected the quality difference of A.crispa.And the results were basically consistent with the principal component and clustering heatmap analysis results.Under this model,sample BLJ3 ranked highest,BLJ1,BLJ2,BLJ6,BLJ7 ranked middle,and BLJ4,BLJ5 ranked lowest.In summary,among the seven indicators,the extract,total saponins,moisture,total ash and acid-insoluble ash content can be used as the main indicators for evaluating the quality of A.crispa from different areas,especially the extract and total saponins content providing the greater quality information.Among the two evaluation models,the CRITIC-grey relational analysis model is more suitable for the comprehensive quality evaluation of A.crispa.Among the seven batches of samples,sample BLJ3 from Pengshui County,Chongqing City is of the best quality,which could be used as the first choice for the high-quality provenances of A.crispa.BLJ1,BLJ2,BLJ6,BLJ7 are of medium quality and can be used as alternative provenances.The quality of the two samples(BLJ4 and BLJ5)from Guiyang,Guizhou Province is poor,which should be avoided as much as possible in the screening of provenances of A.crispa.