Hi Leonardo,

Te Hurst coefficient refers to the fractality or fractional aspect of a series, that is, its long-range dependence. I suggest you to check out ARFIMA models, in which a parameter "d" captures the long-memory component of the series.

Look to your data. As a long term practitioner your aspirations for a model are good but the data will always let you down. The model is a black-box the data the real world so if you try to apply conclusions from errant data what will you get? Just be sure you wrap all you conclusions in good warranties.

Hi Hector,

 Yes, I used AR1MA (and its variants) in order to analyze memory, but, in addition, I am interested in use differents  Hurst coefficients, because it indicates or it can be related to a chaos-complexity indicator. 

Thanks a lot!

Greetings.

Dear Stefan,

Yes, you are right!

 I am studying the prices of agricultural products sold in my town (https://www.researchgate.net/publication/313022947_Prices_of_agricultural_products_in_Santander_Colombia_between_20120616_and_20161230)  in order to cross the historical behavioral with nature and political events.

I have a big question about that: Are the prices efficient?

Does it have memory?

Is there any possibility about forecasting a future price?

Thanks for your advice!

Data Prices of agricultural products in Santander Colombia betwee...

I assume you've seen this https://robotwealth.com/demystifying-the-hurst-exponent-part-1/

Run it in Python and see what you get from Part 2.

I have some good connections in agricultural insurance and would be happy to connect you to them but beware now I know what data you're using my previous comment means even more!

Dear Stefan,

Thanks a lot! Yes, I saw that post before, but it is time to refresh hehe hehe.

I know that the data is'not big enough, but, the collecting is difficult (I read the weekly report and build the dataset)

I would be happy if I can be part of your connections (and extended connection)

Greetings!