Financial Risk Early Warning Analysis Of Pharmaceutical Manufacturing Industry Based On Lstm Neural Network Model

In June 2020,the State Council issued the notice on the key tasks of deepening the reform of the medical and health system in 2021.The policy further promoted the reform of medical security and pointed out the development direction for the future medical industry.With the deepening of the reform,the state gradually launched a more detailed reform plan,further improved the reform framework of medical security,and brought new opportunities to pharmaceutical enterprises,But at the same time,it also exposed a series of problems existing in pharmaceutical enterprises.These problems are particularly obvious in financial management.The unreasonable allocation of funds and resources has affected the long-term and stable development of pharmaceutical manufacturing enterprises.Therefore,it is urgent to build a financial risk early warning model for pharmaceutical manufacturing industry,realize common financial risk prevention and control,and promote enterprises to improve the level of financial risk prevention and control.This thesis adopts the method of combining case and empirical analysis,takes pharmaceutical manufacturing enterprises as the research object,screens the financial indicators in combination with the characteristics of the industry,extracts the principal components by using the principal component analysis method,and establishes the LSTM neural network early warning model suitable for pharmaceutical manufacturing industry,in order to further improve the financial risk early warning research of the industry.Referring to a large number of relevant studies,this thesis introduces the concept,causes and theoretical basis of financial risk,and expounds in detail the basic principle and process of LSTM neural network financial early warning method and its application advantages in enterprise financial early warning.Secondly,it introduces the current situation and industry characteristics of pharmaceutical manufacturing enterprises,analyzes the causes of financial risk of pharmaceutical manufacturing enterprises and the necessity of financial risk early warning.Thirdly,based on the principles of scientificity,practicability,sensitivity and economy,this thesis constructs a financial risk early warning model suitable for pharmaceutical manufacturing enterprises.The specific steps are as follows: the first step is to remove the early warning indicators with high correlation,and finally get 16 financial early warning indicators of 176 listed companies in pharmaceutical manufacturing industry as research samples;The second step is to reduce the dimension of 16 financial early warning indicators of pharmaceutical manufacturing industry by principal component analysis and determine 5 factors;The third step is to set the relevant parameters of the hidden layer,input layer and output layer of the model,and establish the LSTM neural network financial risk early warning model suitable for the pharmaceutical manufacturing industry.Then the model is used to predict the financial risk status of Yibai pharmaceutical in 2021,and it is found that Yibai pharmaceutical has slight financial risk in 2021.Then,the financial risk of Yibai pharmaceutical is analyzed from the four aspects of solvency,operation ability,profitability and development ability,and the corresponding solutions are put forward: for profitability,it is necessary to promote product transformation and enhance the competitiveness of enterprises;In view of the operating capacity,it is necessary to improve the sales mode and strengthen the management of accounts receivable and inventory;For development capability,we should increase investment in innovation and R & D and promote the improvement of core development capability;In view of the solvency,it is necessary to optimize the capital structure of pharmaceutical enterprises and reasonably carry out strategic expansion.In addition,enterprises should constantly improve the governance structure and establish a sound financial early warning system.The conclusions of this thesis are as follows: firstly,the correct percentage of LSTM neural network model constructed in this thesis is 93% in the training set and 81% in the test set,which shows that LSTM neural network model can effectively warn the financial risk of pharmaceutical manufacturing industry;Secondly,the selection of different parameters may affect the accuracy of the model.For example,the accuracy of the model is also different when different layers of hidden layers are selected;Thirdly,the screening and optimization of financial early warning indicators can improve the accuracy of the model.Principal component analysis reduces the correlation between financial indicators by extracting principal components,which greatly improves the accuracy of the model.Finally,this thesis puts forward two application suggestions: the first point is that adding non-listed company data can expand the sample size of the research,and adding non-financial indicators can broaden the dimension of enterprise early warning indicators,so as to effectively improve the accuracy of LSTM neural network model;The second point is to apply the latest RNN model,which can optimize the early warning results of the model.