Machine learning in trading: theory, models, practice and algo-trading - page 1509

New comment
 

There are several problems:

1. it is not clear how to determine the initial initialization parameters

2. Determining whether the state corresponds to the action, only by experience? There must be some objective criteria. We need to do some research,

The pictures look interesting.

If you highlight the signals and look at the resulting balance.

sig <- ifelse(prStat == 1 L, 1, -1)               
cbind((pr$close - pr$open), c(NA, NA,NA, lag(sig))) ->.; na.omit(.)->.; tail(., 432) -> prs
bal <- cumsum(prs[ ,1] * prs[ ,2])
plot(bal, t = "l")

balance

Good luck

 
Maxim Dmitrievsky:
well, I told you so :( psychologists demand DNA extraction from fossilized remains, as ancient as d. dinosaur

Anything is possible if you want it, if you have the right biomaterial.

And about the lecture - it does not reveal at all the declared topic "What is the mind?", in fact there is a declaration of Charles Darwin's ideas, which is not new.

And I extremely disliked the idea that man thinks over his every action with the neocortex, which consumes an excessive amount of energy. No, on the contrary, a man tries to save energy, and thinks about individual tasks, while not liking to think twice and using the decisions he's already made, in order to save energy.

 
Maxim Dmitrievsky:

It doesn't reveal it to you, and you can talk about it endlessly, along with what's-his-name... who is also a psychologist... Peter or whatever... who writes useless visual libs.

Because you can't think logically, without the help of psychological books for morons. Female human beings think more logically and intelligently.

To make the brain work - you need to eat meat, vitamins and solve at least simple mathematical problems.

The second man takes offense at any word... He should be enrolled in a boxing club, so he can be beaten every day... ...then maybe he won't be such a wench :D

Maxim, is it productive to try to show confidence in the knowledge of the subject at the expense of primitive aggressive behavior?

No, it's more a demonstration of weakness and unwillingness to defend your position because of your meager knowledge of the topic.

However, it is not surprising, since this video has brought you a lot of knowledge, there is something to strive for...

In general, let's not talk nonsense and insult people, in this thread about the Defense Ministry and so much rot, already disgusted with inadequate statements.

 
Maxim Dmitrievsky:

I purposely threw the clip to have fun, i.e. I knew what the reaction would be.

And from the very beginning I specifically wrote about psychologists. And you're buying it.

The video is cool in any case, as well as the lecturer.

If your actions are aimed at creating a situation where you can insult others in order to strengthen your masculinity and raise your self-esteem, then that's another story...

In general it's not clear what psychology has to do with it, the lecturer's topic is not fully relevant and therefore there is nothing much to discuss.

 
Maxim Dmitrievsky:

Well, you're falling for it, so I've reached my goal.

You wanted to talk to me so badly but didn't know how to say it or find a reason?

 

Recently we were driving with an acquaintance and he saw a sign like this for the first time in his life. And of course he immediately "understood" what it meant, he has been married for a long time. So that's my point, I thought it was very similar to the situation with the recognition of market patterns and their trading.

Imagine a neural network is trained to recognize different l signs/tables. Everything is normal, it easily recognizes what is in the training sample. And then bang - a new table, and that's it.This is the first time the person has seen it, but he has life experience and a lot of additional knowledge and he immediately "guessed" what it means. So you want a neural network to make money at markets, it must be able to "guess" what "patterns" mean otherwise - no way.You have been trained on the pattern-situations, but a neuronet must grasp the principle, not just memorize the suggested patterns. The market is always different, but the construction principle never changes, like the principle of "drawing" signs under the road signs.If you guess what the next market sign means, you will profit. To look through the market history and search for the similar situation is useless. It is necessary to "understand" and "guess". As it is useless to look through a book of traffic rules of 198 years in search of a "pattern". There were no such signs then. But there were a lot of others and having studied them, you have to learn to guess the meaning of new ones.

 
 

Wizard2018:


There were no such signs back then. But there were a bunch of others, and after studying them, you have to learn to guess the meaning of the new ones.

IMHO the market does not have to warn anyone what will happen next, think about why would it be? Large acts or unpredictable, or contrary to predictions of small and medium, and small and medium "do not shoot away" what is happening, they are at the mercy of illusions, as addicts "see" on the carpet all sorts of animals, or even global cosmic events, sometimes it is purely by chance coincides with reality, but most often not. There are no "tablets."

 
Maxim Dmitrievsky:

Impressed, pampered psychologists better not watch


I take it that's a shot in the dark? Well, uh... let's see.

 
Wizard2018:

I was driving with an acquaintance recently and he saw a sign like this for the first time in his life. And of course he "understood" at once what it meant, he's been married for a long time. So what I mean is, I thought it was very similar to the situation with market pattern recognition and trading.

Imagine a neural network is trained to recognize different l signs/tables. Everything is normal, it easily recognizes what is in the training sample. And then bang - a new table, and that's it.This is the first time the person has seen it, but he has life experience and a lot of additional knowledge and he immediately "guessed" what it means. So you want a neural network to make money at markets, it must be able to "guess" what "patterns" mean otherwise - no way.You have been trained on the pattern-situations, but a neuronet must grasp the principle, not just memorize the suggested patterns. The market is always different, but the construction principle never changes, like the principle of "drawing" signs under the road signs.If you guess what the next market sign means, you will profit. To look through the market history and search for the similar situation is useless. It is necessary to "understand" and "guess". Just as it is useless to leaf through a book of traffic rules in 198, looking for a "pattern". There were no such signs then. But there were plenty of others and by studying them, you have to learn to guess the meaning of new ones.

Oh, what a bunch of people. Are you the Trickster I'm talking about? Note that I capitalized it :-)

1...150215031504150515061507150815091510151115121513151415151516...3495
New comment
Reason: