Machine learning in trading: theory, models, practice and algo-trading - page 1549

[Maxim Dmitrievsky]

Ilnur Khasanov:
A long time ago, would have started streaming, on YouTube, on Twitch. Will you?

About suffering on the subject of MO in trading? )

If someone is interested, it is possible, while I am doing something, you can stream or record

Do you hear the sound from the laptop microphone? I need a studio microphone.

[Ilnur Khasanov]

Maxim Dmitrievsky:

About suffering on the subject of MO in trading? )

If someone is interested, it is possible, while I am doing something, you can stream or record

Do you hear the sound from the laptop microphone? Need a studio microphone

Not necessarily a specific topic. Each stream - a new topic) - now they do so) Just also work as before, but in parallel online, on Twitch there is a cool category for this - science and technology.
I am interested, and even very. Yes, micro is needed, there are noises.

[Maxim Dmitrievsky]

Ilnur Khasanov:
Not necessarily a specific topic. Just work like before, but in parallel online, there is a cool category for it on Twitch - science and technology.
I am interested, and even very. Yes, micro is needed, there are noises.

Yes, I think I will, as soon as I get a normal computer for that

I'm hooked on someone else's streams myself).

[Maxim Dmitrievsky]

1

[Maxim Dmitrievsky]

In the next part, the sound will be studio sound.

[Кеша Рутов]

Maxim Dmitrievsky:

In the next part, the sound will be studio sound.

I approve of your streams. You are welcome to continue. Just scale to code, badly visible on the laptop and do not aaa-cite, take an example from Vladimir Vladimirovich, he does not aaa-cite, it went out of fashion a long time ago.

[Igor Makanu]

Kesha Rutov:

I approve of your streamers. You are welcome to continue. Just scale up to code, it's hard to see on a laptop and don't aaa-cite, take an example from Vladimir Vladimirovich, he doesn't aaa-cite, it went out of fashion a long time ago.

who is Vladimir Vladimirovich?

[Roffild]

Example of using MLflow to store and compare results. The tester screenshot has to be done manually.

import mlflow
from roffild.mqlport import *
mlflow.set_tracking_uri("file:///" + str(pathlib.Path(PATHFILESCOMMON, "mlruns")))
params = [
    'params.VarOpen2',
    'params.VarClose',
    'params.VarOpen_Points',
    'params.VarTime',
    'params.Period_1',
    'params.Period_2',
    'params.TakeProfit_Multi',
    'params.StopLoss_Multi',
]
filter_string = ""
filter_string = "params.VarOpen2='52'"
for index, run in mlflow.search_runs(filter_string=filter_string).iterrows():
    #print(str(run["run_id"]))
    #print(run.keys())
    filename = str(run["run_id"])
    for p in params:
        filename += "_" + p.replace("params.", "") + "=" + str(run[p])
    filename = f"result_{filename}.png"
    print(filename)
    for glb in pathlib.Path(str(run["artifact_uri"]).replace("file:///", "")).glob("*.PNG"):
        print(glb)
        pathlib.Path(PATHFILESCOMMON, filename).write_bytes(glb.read_bytes())

The result is screenshots of the tester with parameters:

result_2d93efb4283748769d34c9a8dcaab155_VarOpen2=52_VarClose=5_VarOpen_Points=35_VarTime=10_Period_1=5_Period_2=20_TakeProfit_Multi=0.5_StopLoss_Multi=1.png
result_ce93896d4f7b4b70bb1d347f5510a85b_VarOpen2=52_VarClose=5_VarOpen_Points=35_VarTime=60_Period_1=5_Period_2=20_TakeProfit_Multi=0.5_StopLoss_Multi=1.png

[Maxim Dmitrievsky]

Roffild:

Example of using MLflow to store and compare results. The tester screenshot has to be done manually.

The result is screenshots of the tester with parameters:

Why not just some tester take azipline for example. What's the advantage?

Then you don't need to run anything in MT5 either, then you don't need the dll either.

[forexman77]

Served just prices to learn the random forest (lagging by one bar). I got a forecast red line. I noticed that when the trend sections, the forecast line does not hit at all.