How to form the input values for the NS correctly. - page 28
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
this is clearly a useless conversation.
---
I've had one forced conversation with a boor in my life--
---
the guy was advanced... in his line of work, but he was a boor
he thought he was smarter than everyone around him - even though he wasn't much different
he would not be polite in conversations
in the canteen with his mouth full, talking loudly about something
always thought he was right about everything.
when he was slowed down, he would push his way in - once he was even punched - by the way, it didn't help for long
When someone was clearly superior to him in his work, he would just go quietly away
but if he ran into a man who was a little ignorant about something, he laughed at him, belittled him, prostrated himself like a rooster in a hen house where there was no other rooster
he spoke disparagingly - through his lip
Well, it's useless to try to encourage such people to behave normally.
Such people don't know how to behave
they think they are doing the right thing but they are doing the right thing within "their" ethical perception of the world - if ethics is even an issue here
One easy way is to ignore them and they become uninterested
.
the second way is to step aside - when they are told they are illiterate
You forgot to tear your tunic, sprinkle your head with ashes :-)
No, I haven't forgotten, I just don't have a tunic, as it's too hot in Tashkent to put anything on. I don't sprinkle ashes on my head because I haven't got lice yet. When I do, I'll be sure to sprinkle it on.
Well enough with the flubbing, Mr. prude. Let's meditate on the subject better.
He often writes beautiful and seemingly correct things, but on closer inspection there is always a catch, something that nullifies all his pompous phrases and statements.
Don't blame the mirror if you don't know what you're doing.
The catch is not because I've made up something or anyone else. It's because markets are non-stationary. So anyone who tries to take a profit with cavalry charges, like interpolations and approximations to history, can easily meet a catch.
Yes, and then my scribble is not the final truth, because, like any other person and not being a sapper I have every right to be mistaken. And therefore I advise to take graffomania not too seriously, but to have my own opinion and to treat everything written not only by me with a sufficient share of criticism (and not criticism). Only I am almost 100% sure that no matter how experienced a trader is, the tricks on financial instruments will catch him or her sooner or later in the most unexpected places. For an experienced trader, the catch is inevitability of drawdown, and for an inexperienced one - inevitability of a margin call. That's the difference
One easy way is to ignore them and they become uninterested
The second way he gets sidetracked is by being pointed out for his illiteracy.
Why then with such donkeyish obstinacy continue to flub in a thread dedicated to neural network inputs about all sorts of ethics and etc. cultures - multitudes? Nobody forced you to talk, you've written yourself that this conversation is useless, so just ignore this thread and go away, if everything is so easy for you, as in hollywood film library.
You Puritans, for fuck's sake!>>" Helpline".
Good, but a bit long...
Then why are you so stubbornly continuing to flub in the thread dedicated to neural network inputs about all sorts of ethics and other culture-multultures? Nobody asked you to talk, you've written yourself that it's useless, so ignore this thread and go away, if everything is so easy for you, as in hollywood film library.
You puritans, for fuck's sake!Well, a boor is a boor.
--
You need to learn how to talk to people and not be rude, even if they're wrong.
You always talk in the style I described in the message that wasn't written to you.
You're always talking through your mouth like you're all assholes.
even though you're not much of a talker yourself.
"We're having fun.
Like a contest--
On the ?????s, whistling,
On ???? knitting."
(ancient ditty)
Of course, the zig-zag has zero informativeness for the network.
as an input finding tool by itself yes! and not just for a neural network
But it is probably good as a tool to find turning points in the nearest history
I'm talking about the learning process and finding pivot points in the nearest history
--
Unless, of course, the network's goal is to detect a reversal or not detect it when it's forming or just forming
then maybe we shouldn't teach the network to look for reversal points
as an entry search tool itself yes! and not just for a neural network
but as a tool to find pivot points in the nearest history is probably a good one
I'm talking about the learning process and finding pivot points in the nearest history
--
Unless, of course, the network's goal is to detect a reversal or not detect it when it's forming or just forming
then maybe you shouldn't teach the network to look for reversal points
In my opinion, for pivot points either, because the zigzag does not give the network information about the price movement between the pivot points. In other words - in a series of zeros it is difficult to predict when the ones will go and vice versa. It is easier for the network to repeat all this with one step lag - "today will be like yesterday, and tomorrow like today". A well-known effect when using a zig-zag. I've had this happen a lot. But on OOS - plum.......
In my opinion, for pivot points as well, because the zigzag does not give the network information about the price movement between pivot points. In other words - in a series of zeros it's hard to predict when the ones will go and vice versa. It is easier for the network to repeat all this with one step lag - "today will be like yesterday, and tomorrow like today". A well-known effect when using a zig-zag. I've had this happen a lot. But on OOS - plum.......
see
in general - it's enough to get at least a simple neural network output like 0 1
where 0 is keep sell 1 keep buy - there will already be a result
or -1 0 1 where -1 is in the Sell, 0 is in Flat, 1 is in Buy
you may use a simple custom indicator, for example on the convergence divergence
and then go from there what to enter and what to learn
---
the most interesting moment is the moment of switching between sell and buy
the grid shouldn't blink and flicker.
that point should be a pivot.
And if you teach it on the nearest history, and this is how the nets are taught - no one would teach the net on the history of 1999
On page 23 of the topic, Yury Reshetov raised an interesting (to me) question:
Let's take the simplest perceptron and connect some indices and their combinations to its input. What gives the largest profit factor on this very perceptron, i.e. the best fit in testing with a constant lot (without MM), will quite probably pass forward tests using a more complex architecture. Why, it is easy to explain. After all, the perceptron is a linear classification. Which means we'll get linear separability on the inputs by pattern. (He cited the MTS Combo as an argument.)
...And non-linearity without linearity is a barefaced fit. So it's a waste of time...
To paraphrase, here is a proof that it is possible to get a linearly separable classification for the market.
Myself, after reading some clever books in the place where it says about impossibility to make an "exclusive or" linear network, I took it as impossibility to use linear networks for market (for that logical reason, that market is much more complicated than a simple "exclusive or" :).
But maybe it is not? Maybe Yuri is right? And no need to razzle-dazzle a bunch of books on non-linear, but just beat everything with planes?