Markeone’s ML Proof of Value, February 2023

markeone
Superalgos | Algorithmic Trading
2 min readFeb 27, 2023

Testing, updates, testing and some more updates.

February has been a busy month for Machine Learning.

Markone has taken the initiative and seems to have cured the ills of the forecasting candles.

Forecasted candles on a test client. Not very accurate yet.

Inspecting the Bitcoin-Factory.json datamine. Markeone spotted that Nuno had made a couple of changes to the code. Changing ‘forcast’ to ‘forecast’ in a couple of places previously. Markeone did a full scan of the bitcoin-factory code and found there were probably more than twenty places were the spelling was incorrect or inconsistent. Markeone made the changes across the code. Setup a test network, server, clients and forecaster. And the above results were achieved! DJ05GaZ was then able to begin testing and began also seeing successful results. So the change was added to the develop branch and steps to bring testnet into line.

With these forecast candles working it’s been mission on to test the data mines within Superalgos, to make sure these will work through the machine learning systems. This work is undertaken by both markeone and DJ05GaZ.

Several datamines have been updated including: Mayer_Multiple, Schaff_Trend_Cycle, Exponentential_Moving_Averages and more.

Testing all the data mines is no small job and some great work by other members of the Machine Learning have seen life brought back toward this part of Superalgos.

Theblockchainarborist did a sterling job fixing up the Testnet test server also adding a new script which sets up the config for a test server and datamine. This has meant adding the new datamine configs is far more streamlined and easy!!

Nuno has also been instrumental in finding issues that were causing some problems in the testnet network. Many thanks have to go to Nuno for seeing the issue which was the script mentioned above had a slight mis-configuration that was easily rectified once seen. Bravo gents!

For those interested in setting up for the ‘no rewards’ testing that is now currently underway. A great place to learn about the project is in the discord channel created for Machine Learning. Come and join us and see where it’s heading.

Lets hope for continued progress.

--

--