Abstract
In this article, we estimated the spectral envelope density function using the modified Daniell Kernel function for a non-stationarity time series, we solved the non-stationary problem by dividing the spectral envelope into a number of local spectral envelopes to reduce the effect of the trend, the time series used in the this article was the sequence of the nitrogenous bases of Escherichia coli bacteria's DNA. The problem of the article is dealing with categorical data in a non-stationary time series.
