论文部分内容阅读
目的:对车载近红外模型库SFDA_Ident(2.6.4)中抗菌抗病毒类化学药粉针剂定性分析模型进行验证并优化。方法:采集国家评价抽验品种注射用阿昔洛韦、更昔洛韦、炎琥宁、穿琥宁共计455批次样品的近红外漫反射光谱,对模型进行验证;在模型中增加注射用喷昔洛韦,并对建模品种、谱段、因子谱等参数进行调整,建立优化模型;同时采集阿昔洛韦、更昔洛韦、利巴韦林、脱水穿心莲内酯琥珀酸半酯、苦参碱的对照品近红外光谱,建立上述粉针剂二级串联模型。结果:模型验证总的正确识别率为62.9%。优化后模型的正确识别率为100%;建立的5个串联模型正确识别率为100%。结论:模型优化后更加科学合理,增加了使用范围,提高了准确率,适合药品现场快速筛查。
Objective: To validate and optimize the qualitative analysis model of antimicrobial and antiviral chemical injection in SFDA_Ident (2.6.4). Methods: The near-infrared diffuse reflectance spectra of a total of 455 batches of acyclovir, ganciclovir, metronidazole, and chuanhuning were collected from the national appraisal sample to verify the model. Penciclovir for injection We also adjusted the parameters of modeling varieties, spectral bands and factor spectra to establish the optimized model. At the same time, acyclovir, ganciclovir, ribavirin, dehydroandrographolide succinate, Alkali reference substance near infrared spectroscopy, to establish the above-mentioned two-stage injection of powder injection model. Results: The overall correct recognition rate of model verification was 62.9%. The correct recognition rate of the optimized model is 100%. The correct identification rate of the five series models established is 100%. Conclusion: The model is more scientific and rational after optimization, which increases the scope of application, improves the accuracy and is suitable for rapid screening of drugs.