Parameter Estimation Technique for general datasets

An estimation procedure for many types of data

In the paper "An estimation technique for Time Indexed gaussian Mixture Models", we propose a model specification that can be used to describe data with spikes, jumps, mean reversion, geometric brownian motion, you name it...

We then develop an estimation procedures to extract the parameters from data generated from this model.

The original intent was to develop a model for electricity price spike, however it can be extended far beyond this.

A more general extension of this basic idea using particle filtering is developed in another paper.