Wavelet Neural Network Predication For Epidemic Outbreak

Density function estimation is a kind of method to solve the sample distribution density function in nonparametric statistics, which is a hot spot in the research of statistics. There are many kinds of density function estimation methods. In this paper, we mainly study the two estimation of density function, and give an application example of the two estimation methods. The following modules are divided into the following modules:First of all, the two estimation problems of density function are estimated by introduction. The background of wavelet neural network is introduced and the research status and the function of wavelet neural network is briefly introduced,including its function, adaptability and the ability to analyze and process information,and make a brief introduction to the history of the development of wavelet. Kernel estimation is a kind of estimation method which is not continuous from the histogram estimation. It analyses the advantage of kernel estimation, and gives a brief introduction to the development history of kernel estimation, and further gives the innovation points of the paper.Secondly, the function of wavelet neural network is introduced, the definition of wavelet function is given. Some commonly used wavelet bases are listed. The wavelet estimation of density function is introduced, and the algorithm and icon of wavelet neural network is given. At last, a numerical example is given to predict the sudden epidemic disease with wavelet neural network, and the corresponding diagrams are made, and the true density curve is compared with the wavelet estimation density curve.Again, This paper introduces the basic knowledge of kernel density estimation, the definition of kernel density estimation, and gives some common kernel functions.The Gauss kernel function to be used to do a brief introduction. At the end of this chapter, we give a numerical example of the prediction of the burst infectious disease by using kernel estimation, and make the correlation diagram, and compare the true density curve with the kernel density estimation.Finally, the two methods of density function estimation are compared..According to the sources and types of data, when the selected data source in positive normal distribution or a mixture of normal, two estimation methods have little difference; When the data is extracted from the mixture of lognormal and normal distribution, the result of wavelet estimation is better than that of kernel estimation,and then choice appropriately wavelet estimation or kernel estimation.