Which approach you use depends on the type of variables you have as mediator(s) and outcome. There are several approaches available, some based on the SEM framework, and others based on the potential outcome framework. I am providing a link below to a paper I recently wrote in which we first describe all the approaches available, and then we compare them using simulation.

I suggest you give this a quick glance, since this should help you determine which approach suits your data the best. At the very least, we provide a very comprehensive bibliography that should point you  to the relevant papers you may want to delve into.

Ariel

Article Using mediation analysis to identify causal mechanisms in di...

I am grateful sir

Please check out my paper: https://www.researchgate.net/publication/255909532_Sparse_model_from_optimal_nonuniform_embedding_of_time_series

It describes building high degree polynomial models. There is also a discussion on OLS.

Article Sparse model from optimal nonuniform embedding of time series