Pazen window estimate MATLAB source code (with sam
no vote
Contents: 1. the distribution of samples. Mvnrnd using MATLAB's function, given the mean and covariance matrix as well as the distribution of samples to obtain samples. The function principle is actually using probability density function of the normal distribution random to generate a series of random data, namely distribution of samples. 2. using Parzen window estimation and KN, neighbor estimator for the results are basically the same. But the use of different practical problems in different context, get the accuracy of different. Should be based on specific practical issues and data properties to determine which method to use. Results: first built the two-dimensional Gaussian dataset, the Parzen window estimation method is then used to make non-parametric estimation of the PDF, and compare them with the original PDF.