Machine learning in trading: theory, models, practice and algo-trading - page 2883
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
That's 3 vertices of a zigzag with anything in between.
и?
и?
Use the last 6 vertices of the ZZ as fiches. And you get the same thing, only simpler.
Use the last 6 vertices of the ZZ as fiches. And get the same thing, only simpler.
it's just an example from scratch, you had to show the concept on something.
What exactly do you want to search for and in what way?
For example we have this pattern, three peaks, or anything (rule, event, pattern, cluster).
It's just an example from scratch, something to show the concept.
elibrarius #:
1) Ну, хорошо. Дальше сложнее... вершины/развороты трендов можно считать событиями.
2) And here's how to come up with other events (which are anything) on the graph in between and then describe them...?
1)
First of all, let's start with terminology...
For me an event is a certain event(tuftology) in the market that is of interest to a trader. It can be anything - tops , trend reversals, price at a level, rebound, time, bid in the stack, large volume, etc.... Anything of interest.
An event can be described
1... functionally
2...can be a log...rule
3... can be a code... if you think about it, 2 and 3 are the same thing.
2)
You don't need to invent anything, there are algorithms that invent rules themselves, or write code...
1... algorithm came up with the rule/code
2...tested it on history
3...get a result
You don't have to invent anything, there are algorithms that make up the rules themselves, or write the code...
I don't work in R or Python.
I sympathise)))
С++ ?
Well, I've given you an algorithm that fits your requirements.
1) without time constraints, as we write what we need ourselves
2) any logic of searching for regularities, because we write what we need ourselves.
3) any choice of pattern description, even by log rules or by functions , because we write what we needourselves.
In my proposed concept
these patterns will be equivalent, and the patterns themselves can be of any complexity.
and no AMO can do that.
and there are "stop" rules, and no AMO can do that either.
It means a general purpose AMO with tabular data at the input.
Don't you have an explicit limit on the pattern length of 10 bars? It is clear that the pattern itself is not tied to time, but the moment of entering a trade is somehow tied to the pattern? And the number of bars by which everything is counted from the moment of entry is clearly limited by the size of the pattern. Still, we get a calculation window of a given size.
Don't you have an explicit restriction on the pattern length of 10 bars? It is clear that the pattern itself is not tied to time, but the moment of entering a trade is somehow tied to the pattern? And the number of bars by which everything is counted from the moment of entry is clearly limited by the size of the pattern. Still, we get a calculation window of a given size.
It combines both...
1) rules can look at the last 10 candlesticks ( rigid indexing [1 : 10] )
2) rules can be searched on all data regardless of the index, it is a loop [i], also these rules can look not only at their index but also at [i+n ].
X[2, "close"] means the second clause from the last price.
X[i, "close"] - this is the cloze in the cycle (can be anywhere even +100)
X[i+5, "close"] - anywhere +5.
Roughly speaking, for example, we can find an independent pattern of three candlesticks, 100+ candlesticks ago and immediately compare it with the penultimate price of the close ...
In other words, the model understands what the last prices are and understands what the prices just were, regardless of when....
If you are interested, I can send you the code, the functions, how the rules are searched, the fitness functions.
Roughly speaking for example we can find an independent pattern of three candlesticks, 100+ candlesticks ago and immediately compare it with the penultimate price of a cloze.
In other words, the model understands what the last prices are and understands what the prices just were, regardless of when....
If you are interested, I can send you the code, the functions, how the rules are searched, the fitness functions.
Drop me a line and I'll take a look. This could be interesting.