Objective:To establish a rapid quantitative model of sodium chloride content in Halitum of traditional Chinese medicine by using X-ray diffraction technology and machine learning algorithm.Methods:The data of X-ray diffraction patterns of 90 batches of Halitum samples were collected,and the rapid prediction models of X-ray diffraction were constructed by using partial least squares(PLS),support vector regression(SVR)long short-term memory(LSTM)according to the reference values determined by the content determination method of Halitum in the Chinese Pharmacopoeia 2020 edition.Results:The data preprocessed by multivariate scatter correction(MSC)and selected by competitive adaptive reweighted sampling(CARS)were better in PLS model and LSTM model.The data pretreated by standard normal transformation(SNV)and selected by CARS feature variable perform well in SVR.Conclusion:The three models show good prediction potential,which shows that the combination of X-ray diffraction technology and machine learning algorithm is feasible for accurate prediction of the content of Halitum in traditional Chinese medicine.

Objective:To establish a rapid quantitative model for the determination of moisture,extractives,and content in Pollen Typhae.Methods:Near-infrared spectra of 91 batches of Pollen Typhae samples were collected.Spectral preprocessing was performed using S-G,MSC,SNV,and CWT methods.Variable selection was conducted using CARS,SPA,and VIP methods,and compared with full-spectrum modeling.Partial least squares(PLS)mod-els were established for the quantitative determination of moisture,total ash,extractives,and content.The model performance was evaluated by calculating the coefficient of determination for the calibration set and validation set(R2 c,R2v),root mean square error of calibration and validation(RMSEc,RMSEv),and residual prediction devia-tion(RPD).Results:The PLS models for moisture,extractives,and content in Pollen Typhae showed R2c and R2v values greater than 0.9,RMSEc and RMSEv values approaching 0,and RPD values greater than 3.Conclusion:In this study,near-infrared spectroscopy was used to construct quantitative prediction models for moisture,extractives,typhaneoside,and isorhamnetin-3-O-neohesperidoside content in Pollen Typhae.This method enables rapid detection of the main quality control indicators of Pollen Typhae,providing strong technical support for its quality supervision.

Objective:To establish a rapid quantitative model of sodium chloride content in Halitum of traditional Chinese medicine by using X-ray diffraction technology and machine learning algorithm.Methods:The data of X-ray diffraction patterns of 90 batches of Halitum samples were collected,and the rapid prediction models of X-ray diffraction were constructed by using partial least squares(PLS),support vector regression(SVR)long short-term memory(LSTM)according to the reference values determined by the content determination method of Halitum in the Chinese Pharmacopoeia 2020 edition.Results:The data preprocessed by multivariate scatter correction(MSC)and selected by competitive adaptive reweighted sampling(CARS)were better in PLS model and LSTM model.The data pretreated by standard normal transformation(SNV)and selected by CARS feature variable perform well in SVR.Conclusion:The three models show good prediction potential,which shows that the combination of X-ray diffraction technology and machine learning algorithm is feasible for accurate prediction of the content of Halitum in traditional Chinese medicine.

Objective:To establish a rapid quantitative model for the determination of moisture,extractives,and content in Pollen Typhae.Methods:Near-infrared spectra of 91 batches of Pollen Typhae samples were collected.Spectral preprocessing was performed using S-G,MSC,SNV,and CWT methods.Variable selection was conducted using CARS,SPA,and VIP methods,and compared with full-spectrum modeling.Partial least squares(PLS)mod-els were established for the quantitative determination of moisture,total ash,extractives,and content.The model performance was evaluated by calculating the coefficient of determination for the calibration set and validation set(R2 c,R2v),root mean square error of calibration and validation(RMSEc,RMSEv),and residual prediction devia-tion(RPD).Results:The PLS models for moisture,extractives,and content in Pollen Typhae showed R2c and R2v values greater than 0.9,RMSEc and RMSEv values approaching 0,and RPD values greater than 3.Conclusion:In this study,near-infrared spectroscopy was used to construct quantitative prediction models for moisture,extractives,typhaneoside,and isorhamnetin-3-O-neohesperidoside content in Pollen Typhae.This method enables rapid detection of the main quality control indicators of Pollen Typhae,providing strong technical support for its quality supervision.

Objective To quantitatively evaluate the structure and content of the long-term care insurance(LTCI)policy in China's deeply aging areas.Methods Using the Policy Modeling Consistency(PMC)index model indicator design method,a LTCI policy evalua-tion index system was constructed,consisting of nine primary indicators and 34 secondary indicators.A total of 123 provincial-level LTCI policies issued in deeply aging regions of China between June 1,2014,and October 1,2024 were analyzed.High-frequency word extraction was performed using ROSTCM 6.0,and a social network diagram of LTCI policies was created.The policy structure and content were quantitatively evaluated and ana-lyzed based on the established policy evaluation index system.Results The main content of LTCI policies in deeply aging areas of China covered services,institutions and assessment.The highest policy score was 7.28,and the lowest was 2.20,with an average score of 5.00.There were 25 perfect policies,63 excellent policies,28 good policies and seven qualified policies.In the dimension of policy content,the indexes of five primary indicators of policy evaluation,policy target groups,policy nature,policy perspective and policy tools were 0.60 or more;while the indexes of four primary indicators of policy content,incentives and constraints,policy timeliness,and policy level were 0.50 or less.Conclusion LTCI policies issued in China's deeply aging areas provide comprehensive coverage in aspects such as poli-cy evaluation,policy target groups and policy nature,and need to be improved in policy tool selection and the construction of incentive and constraint mechanisms.

Objective To quantitatively evaluate the structure and content of the long-term care insurance(LTCI)policy in China's deeply aging areas.Methods Using the Policy Modeling Consistency(PMC)index model indicator design method,a LTCI policy evalua-tion index system was constructed,consisting of nine primary indicators and 34 secondary indicators.A total of 123 provincial-level LTCI policies issued in deeply aging regions of China between June 1,2014,and October 1,2024 were analyzed.High-frequency word extraction was performed using ROSTCM 6.0,and a social network diagram of LTCI policies was created.The policy structure and content were quantitatively evaluated and ana-lyzed based on the established policy evaluation index system.Results The main content of LTCI policies in deeply aging areas of China covered services,institutions and assessment.The highest policy score was 7.28,and the lowest was 2.20,with an average score of 5.00.There were 25 perfect policies,63 excellent policies,28 good policies and seven qualified policies.In the dimension of policy content,the indexes of five primary indicators of policy evaluation,policy target groups,policy nature,policy perspective and policy tools were 0.60 or more;while the indexes of four primary indicators of policy content,incentives and constraints,policy timeliness,and policy level were 0.50 or less.Conclusion LTCI policies issued in China's deeply aging areas provide comprehensive coverage in aspects such as poli-cy evaluation,policy target groups and policy nature,and need to be improved in policy tool selection and the construction of incentive and constraint mechanisms.