Study On Profitability Evaluation Of The Chinese Bio-pharmaceutical Listed Companies

Bio-pharmaceutical industry in China is developing rapidly in recent years. By the year of 2014, the output value of Chinese bio-pharmaceutical industry has been more than 2 tril ion yuan. Behind the rapid development of bio-pharmaceutical industry market, there is a huge demand for bio-pharmaceutical product innovation. However,the bio-pharmaceut- ical industry is an industry with high risk and high income.The evaluation of the profitability of the enterprise is of great significance to investors to make correct investment decisions.It is not only do help for investors to put resources into the project with good return,but also can help investors to identify the risk,reducing the investment loss.This paper is regarding the listed companies with bio-pharmaceutical in China as the research objects. Selecting all the pharmaceutical biotechnology sector listing corporations which is classified by SWS as research samples.Then divid the samples into training samples and testing samples.And construct the quantitative model of profitability evaluation. During the course of this study,we divided the samples into four levels according to the standard of the annual net profit excluding non recurring gains and losses after the asset scale of the year T. Then based on the financial report of the year T-1,and the profit situation in T based on the financial data, use the method of principal component analysis and the BP neural network to process the data and train samples.When the training error is up to the preset accuracy, save the trained neural network model.Because of the model has the ability of learning and memory,it has remembered the financial characteristics of different levels of profitability of the enterprise.Then input the financial report of current year,we can get the profit level of the next year.The study found that using the methods of PCA and neural network to evaluate the profit level of the pharmaceutical listed companies, can effectively reflect its advantages in data processing and analysis.Finally the testing result of the model show that the accuracy for different levels of profitability of enterprises is nearly 80%. The effect of the model is significant.