DOC: Expand docstring of ExponentialSmoothing model with parameters

[pierre-haessig]

Hello,

Here is a PR which expands the docstring of ExponentialSmoothing, in the spirit of what was already there for other models like SARIMA or UnobservedComponents.

In particular:

• recall ETS model equation, to define the parameters
• explain the (hyper)parameters and give their name in the code

Also, I dived in a bit in how the state space is implemented, but not to the point of yielding the complete state space equations, because I don't know how useful it can be. On the other hand, describing the state vector may be useful for people trying to exploit the state vector. But perhaps it would be better to go "all-in" and write down the state space equations?

Finally, there are minor changes such as homogenize the references formatting.

Anyway, I hope it is useful and that I didn't write major mistakes. This text is based on my analysis of the code in src/models/exponential_smoothing.jl and comparing with the Forecasting: Principles and Practice, the Pythonic Way book.