Dear Jiří Kroc : Yes I have data and many identified attributes to categorise the data but still predictions are not as expected hence found some attributes are missing which need to be considered.
Here if someone has expertise in these stuffs, they can share their views which may be helpful.
There is always a way to construct a reliable description of any data structure that evolves with time. It requires several steps.
- building a database
- let know your data in depth
- scan literature for known approaches
- learn how to correctly use entropy
- construct advanced data extraction approaches
- compare the predictions to real data
It is a heck of work. Currently, I am working on the final version of one paper of this type that predicts arrhythmias from raw ECG recordings. It is written in a way that explains principles to everyone in any filed.
By the way, to explain all in one paper is extremely difficult and time-consuming but once you know the principles, you can go fast forward.
Hopefully, the paper gets finished in a quarter of a year.