Research On Financial Early Warning For Pharmaceutical Manufacturing Companies Based On Machine Learning

In recent years,with the aging of China’s population,the promulgation of the national three-child policy,and the sudden outbreak of the new crown pneumonia epidemic,people have devoted more attention to their health,and the demand for pharmaceutical products has increased while the requirements for pharmaceutical product quality and the expectations for pharmaceutical manufacturing companies listed on the stock exchange have become higher and higher,and pharmaceutical manufacturing companies have played an increasingly important role in our daily lives.The changes in the external environment,such as the national health insurance cost control policy,the promulgation of GMP guidelines,and the entry of joint ventures,have brought rapid development to pharmaceutical manufacturing companies in China,while also putting increasing pressure on the financial risk management of pharmaceutical companies in China.The pharmaceutical manufacturing industry has a much higher financial risk than other industries due to its long R&D cycle,high R&D risk,and high capital requirements.The design and application of financial risk early warning models can help pharmaceutical companies to identify risks as early as possible and protect their smooth development.Therefore,it is important to establish a financial early warning model suitable for China’s pharmaceutical manufacturing industry for the healthy and stable development of the industry.The main research objective of this thesis is to design a financial risk early warning model applicable to China’s pharmaceutical manufacturing sector based on machine learning algorithms.Firstly,based on reading and organizing a large amount of relevant literature,this paper describes the current situation of financial risk,early warning models and financial early warning related research in pharmaceutical manufacturing industry at home and abroad,followed by the definition and causes of financial risk,the definition of financial risk early warning and the principles,advantages and disadvantages of five classical models of financial risk early warning.Next,the industry characteristics of pharmaceutical manufacturing industry and the characteristics of financial risks faced by this industry are analyzed with the author’s internship experience.Subsequently,80 pharmaceutical listed companies with similar asset size in domestic A-shares from 2016 to 2021 were selected as research objects according to the pairing principle,and the PCA technique was applied to downscale the 30 sample indicators initially selected,condense them into 7 representative indicators and compose the indicator system.Then,the sample data were divided into training group and validation group,and the SVM model with RBF as kernel function and a 7-9-2 BP neural network model were constructed by importing the training group data in MATLAB software.Based on this,the two models were validated using the data of the divided 20 validation groups.The results show that the BP neural network model trained in this paper achieves 95% prediction accuracy in financial risk warning for pharmaceutical manufacturing companies in China,while the SVM financial risk warning model with RBF as the kernel function achieves 90% prediction accuracy.By comparing the two models it was found that the BP neural network has higher prediction accuracy and lower error rate in this field.Finally,the trained model was applied to HY case company,and the real data of HY company was input into the trained financial risk early warning model for simulation experiments,and the accuracy of the model was verified twice by comparing the financial data with the development history and current situation of HY company,and the potential financial risks of the company were found and the corresponding improvement measures were proposed.