基于神经网络和实物期权的新药研发项目估值模型研究  

Valuing model of new drug R&D project based on neural network and real options

作　　者：薛永刚[1] 张明丽[1] 

机构地区：[1]广东药学院医药商学院,广州510006

出　　处：《中国新药杂志》2016年第4期370-375,383,共7页Chinese Journal of New Drugs

基　　金：2013年广东省医学科研基金:实物期权在新药研发项目投资决策中应用研究(A2013311)

摘　　要：新药研发项目具有周期长、高风险的特征,本文研究实物期权在新药研发项目估值评估中的应用方法和效果。首先通过神经网络模型预测项目研发周期内无风险收益率的变化过程,然后在变化的无风险收益率条件下,通过二叉树期权模型计算项目资产价值变化过程和期权价值,最后与传统的净现值法作对比分析。实证研究表明:1由于考虑了项目决策过程中的灵活性,实物期权可以更加客观准确估计新药研发项目价值。2在周期较长的研发项目评估中,应该考虑无风险收益率的变化过程,可以使评价结果更加准确合理。3神经网络模型在无风险收益率预测中具有良好的性能,能够满足项目风险评估的需求。The project of new drug R＆D has a great many uncertainties and longer period. In this paper, we investigated the application of the real option in valuing new drug R＆D. First, the risk-free return rate in the period of the R＆D project was estimated based on the neural network ; then the value of the assets and the real op- tion were calculated based on the binomial tree option pricing model. The results were contrasted with the tradition- al net present value. The empirical results have shown the following conclusions. （1） Considering to the flexibility in the decision-making process, the real option model can estimate the uncertainty of the new drug R＆D project. （2） The variety of risk-free return rate should take into account in valuing risk of project with a long period, which can improve the veracity of the estimating value. （3） The neural network has high performance in forecasting the risk-free return rate, which can satisfy the requirement of valuing the project risk.

关 键 词：新药研发 估值模型 实物期权 神经网络 

分 类 号：R95[医药卫生—药学]