Analysis of the Real-World Application Laws for Zhengan Xifeng Decoction in the Treatment of Hypertension Based on Machine Learning

AbstractHypertension is a common cardiovascular disease. Zhengan Xifeng Decoction (ZGXFD), a classic prescription for adjuvant treatment of hypertension, but its clinical application characteristics and biological information have not been comprehensively analysed. This study is based on real-world data from 7571 electronic medical records of hypertension patients treated by ZGXFD. The Apriori algorithm was used to obtain the coupled herbs of ZGXFD. A convolutional neural network was designed to measure the dose characteristic information of herbs. The topological features of the protein‒protein interaction data were used to analyse the biological information of herbs. The K-nearest neighbour model integrates the above characteristics of herbs into the same framework to observe the composition adjustment laws and mechanism of ZGXFD from multiple dimensions. Eighty-seven coupled herbs with dose characteristics were obtained. The results revealed that ZXGFD regulates cytokines and reduces the inflammatory response and metabolic disorder to achieve the purpose of adjuvant therapy. Moreover, machine learning model is used to analyze real-world data that include clinical and molecular biological data with hierarchical characteristics, which provides a micro-biological explanation for the clinical application of herbs.