基于极限学习机的实物期权定价模型研究-产业经济学专业论文.docxVIP

AbstractWith Abstract With the continuous development of economic globalization，the business and management methods of enterprises are evolving，resulting in increasing and innovated enterprise value types．The enterprises are faced with more future uncertainties and the management flexibility becomes increasingly prominent．Thus，the enterprise valuation plays a crucial role in modem mergers and acquisitions，selling，joint ventures and investment decisions．Traditional methods can not be used to estimate the implicit value of enterprises because of the uncertainties．Real option method can be used to estimate the implicit value．According to the real option，the most part of the enterprise value comes from the decision-making opportunities arising from future uncertain changes．The value generated by these future opportunities is the real option value，which constitutes a major portion of enterprise value． Black-Scholes model is a more popular pricing tool for enterprise evaluation． However，since the premise conditions of Black-Scholes model are very strict and there are subjectivity and fuzziness in Black—Scholes model，the value can not be assessed accurately and efficiently，which makes the model much restricted in practice．Although many scholars have improved and optimized the real option method by using a variety of algorithms to solve the above problems inherent in real option，and the accuracy of real option pricing model has been greatly promoted，the optimal results are still not reached due to the shortcomings of BP neural network．Therefore，1 will introduce ELM(Extreme Learning Machine，ELM)algorithm to improve the real option model in this paper．ELM is more greatly improves the training speed and prediction accuracy．This paper attempts tO simulate sample data of listed enterprises by Matlab for the empirical analysis．Compared to that of improved model based on BP，the result of real option pricing model based on ELM is more effective．So far，we have proved the effectiven