About the Multiplicative Inverse of a Non-Zero-Mean Gaussian Process
By: Marco Lanucara
Potential Business Impact:
Finds patterns in noisy signals.
We study the spectral properties of a stochastic process obtained by multiplicative inversion of a non-zero-mean Gaussian process. We show that its autocorrelation and power spectrum exist for most regular processes, and we find a convergent series expansion of the autocorrelation function in powers of the ratio between mean and standard deviation of the underlying Gaussian process. We apply the results to two sample processes, and we validate the theoretical results with simulations based on standard signal processing techniques.
Similar Papers
Filtering of periodically correlated processes
Statistics Theory
Finds best guesses for hidden signals in noisy data.
On extremes for Gaussian subordination
Probability
Predicts rare, extreme events in complex systems.
Spectral estimation for high-dimensional linear processes
Statistics Theory
Helps understand complex stock market patterns.