Research On Convertible Bond Pricing Based On Kalman Filter And Neural Network

This paper studies the pricing of convertible bonds based on kalman filter and neural network.Firstly,this paper introduces the research background and significance of the subject,the development status of convertible bond pricing at home and abroad,and the relevant research of kalman filter and neural network.In this paper,a pricing model of convertible bonds based on neural network is designed.Compared with the traditional model,this model has certain accuracy improvement.Network,however,such problems as slow learning speed and generalization ability in limiting the use of neural network,a lot of noise samples will lead to a wrong learning as a result,the neural network failure,the actual pricing data often contain a lot of uncertainty,so this article and put forward the combination of the kalman filter and neural network model,this paper do a lot of experiments on the improved method,comparing with the traditional method,good to solve the above problems.The combination of kalman filter and neural network is one of the hot topics in the fusion of various information technologies.The combination of these two methods enables the algorithm to improve the relevant defects of the original algorithm,enhance the practicability of the algorithm,and also improve the accuracy in practical applications,such as the pricing of convertible bonds in this paper.In this paper,we use MATLAB software to simulate the sample data of convertible bonds collected from wind database,snowball finance and Dongfang wealth choice data,and compare the simulation results of the pricing model of convertible bonds under the neural network and the improved pricing model of convertible bonds under the neural network.Experimental results show that the output error and time of the improved neural network are significantly lower than those of B-S and neural network in the pricing of convertible bonds,so the improved pricing result of convertible bonds combined with Kalman filter is more effective.This is of great significance to the study of the value evaluation of convertible bonds and the decision-making of project investment.The main innovation of this paper is to simplify the structure of neural network and apply the combined model of kalman filter and neural network to the pricing of convertible bonds.