Research On Clustering Method Based On Data Enhancement For Analyzing TCM Data

The efficacy and efficacy of traditional Chinese medicine mainly depend on the time,dosage and ratio of medication.Among them,the time to re-take the drug is particularly important because it is directly related to the effectiveness of clinical application.However,the medication guidance of many Chinese medicines often copies ancient medical books and lacks the support of modern scientific analysis methods,which leads to the lack of scientific and practical nature of many medication schemes.In this context,taking the blood concentration data of Dachengqi decoction as an example,in view of the problem of small sample size of traditional Chinese medicine data,this experiment proposes a new method to enhance the data of traditional Chinese medicine use.The data were analyzed based on OPTICS clustering combined with drug half-life,and the re-taking time of Dachengqi decoction was discussed.In this experiment,the following research work was carried out:(1)XGBoost is a gradient boosted tree model with strong interpretability.At the same time,the TCM data in this experimental study are often continuous time series data,and the long-term dependence between the data needs to be considered.The GRU model is able to capture this long-term dependent association,so it can more effectively grasp the temporal correlation in the TCM drug usage data.In order to further improve the accuracy of the model,this experiment proposes to combine GRU and XGBoost models using the inverse distance weighting method.The accuracy of this combinatorial model is significantly higher than that of the single data augmentation model,and the combinatorial model can effectively increase the amount of sample data,so as to more accurately simulate the medication data of traditional Chinese medicines.(2)Generative adversarial networks(WGAN)are able to learn data distributions to generate new data similar to the original data.In this experiment,a generative adversarial network(WGAN)and Gated Recurrent Unit(GRU)in deep learning are used to combine the characteristics of traditional Chinese medicine data to propose a WGAN data augmentation method based on GRU.This method uses GRU to learn the time series features of TCM data,and uses WGAN to generate more TCM data,thereby amplifying the size of the dataset.In order to verify the effectiveness of this method,a series of experiments were designed and the experimental results were analyzed and discussed in detail.Experimental results show that the WGAN method based on GRU proposed in this study can effectively enhance the size of the TCM dataset and improve the accuracy and robustness of the model.(3)In view of the characteristics of TCM medication taking data,this study proposes a clustering method to analyze the re-taking time of TCM.By using the clustering method to divide the medication data into several categories,combined with the concept of drug half-life,the key nodes in the Chinese medicine medication data were identified.In the study of the experimental data of Dachengqi decoction,the key node of the re-taking time was obtained through calculation and analysis.The results of this study provide a scientific analysis method for the re-taking time of Chinese medicine,making the taking of Chinese medicine more scientific and effective.This study can provide certain guiding significance for the clinical application of traditional Chinese medicine.It not only proposes two effective data enhancement methods,but also provides scientific analysis methods for the re-taking time of traditional Chinese medicines through cluster analysis and calculation of drug half-life,and provides certain practical guidance for the clinical application of traditional Chinese medicines.