Probability Density Estimation Based on Nonparametric Local Kernel Regression

摘要

In this research, a local kernel regression method was proposed to improve the computational efficiency after analyzing the kernel weights of the non-parametric kernel regression. Based on the correlation between the distribution function and the probability density function, together with the nonparametric local kernel regression we developed a new probability density estimation method. With the proper setting of the sparse factor, the number of the kernels involved in the kernel smooth was controlled, and the density was estimated with highly fitness and smoothness. According to the simulations, we can see that the proposed method shows a very well performance both in the accuracy and the efficiency.