Recently, I have been working upon a project for a company to forecast its order. Financial data is a mess. It has always been. But as Tou said that there is always order in chaos. We, data scientists, should find order in chaos. The only order I could find is to tame the data. Data is exponential. Data is our love. Data is elixir of Machine learning.
This project was fun. Creating variable transition matrix for different 't' has given me a commendable accuracy of around 97%. Though I am not at liberty to discuss the project due to confidentiality issues.
I'd like to know that has anyone ever worked on variable transition matrix? What was your experience?