Machine learning in trading: theory, models, practice and algo-trading - page 1892
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
Yes! , what is the point of going through all types of AMO if the "backlash" in quality of recognition between them all is less than 5%. They all obviously lack information (features) for deeper knowledge (better classification) of the object, so yes, I started to work exclusively on features and ways of presenting information.
By the way there is an interesting package in python on automatic generation of featuresfeaturetools, I unfortunately did not manage to run it in R-ka, some problems with python at me))) Take a look at it, I think it's an interesting thing.
And what type of predictors have you added?
I haven't gotten to python and R yet - very little time :(
I'm wondering what kind of predictors I can come up with with a regression channel? I have a coefficient, the number of repetitions of channel construction vector, fixing the points of price crossing the borders of the channel.
And, may be, who knows how the regression channel in MT5 is calculated, if its end point is stretched beyond the current date, i.e. into the future?
Anything is possible )
featuretools, unfortunately I never managed to run it in R-ka, I have some problems with python)) Take a look at it, I think it's an interesting thing.
think or proven? what's interesting about it?
In [12]: feature_matrix_customers Out[12]: zip_code COUNT(sessions) NUM_UNIQUE(sessions.device) MODE(sessions.device) SUM(transactions.amount) STD(transactions.amount) MAX(transactions.amount) SKEW(transactions.amount) MIN(transactions.amount) MEAN(transactions.amount) COUNT(transactions) NUM_UNIQUE(transactions.product_id ) MODE(transactions.product_id) DAY(date_of_birth) DAY(join_date) YEAR(date_of_birth) YEAR(join_date) MONTH(date_of_birth) MONTH(join_date) WEEKDAY(date_of_birth) WEEKDAY(join_date) SUM(sessions.SKEW(transactions.amount)) SUM(sessions.NUM_UNIQUE(transactions.product_id)) SUM(sessions.MAX(transactions.amount)) SUM(sessions.MIN(transactions.amount)) SUM(sessions.STD(transactions.amount)) SUM(sessions.MEAN(transactions.amount)) STD(sessions.SKEW(transactions.amount)) STD(sessions.NUM_UNIQUE(transactions.product_id)) STD(sessions.MAX(transactions.amount)) STD(sessions.SUM(transactions.amount)) STD(sessions.COUNT(transactions)) STD(sessions.MIN(transactions.amount)) STD(sessions.MEAN(transactions.amount)) MAX(sessions.SKEW(transactions.amount)) MAX(sessions.NUM_UNIQUE(transactions.product_id)) MAX(sessions.SUM(transactions.amount)) MAX(sessions.COUNT(transactions)) MAX(sessions.MIN(transactions.amount)) MAX(sessions.STD(transactions.amount)) MAX(sessions.MEAN(transactions.amount)) SKEW(sessions.NUM_UNIQUE(transactions.product_id)) SKEW(sessions.MAX(transactions.amount)) SKEW(sessions.SUM(transactions.amount)) SKEW(sessions.COUNT(transactions)) SKEW(sessions.MIN(transactions.amount)) SKEW(sessions.STD(transactions.amount)) SKEW(sessions.MEAN(transactions.amount)) MIN(sessions.SKEW(transactions.amount)) MIN(sessions.NUM_UNIQUE(transactions.product_id)) MIN(sessions.MAX(transactions.amount)) MIN(sessions.SUM(transactions.amount)) MIN(sessions.COUNT(transactions)) MIN(sessions.STD(transactions.amount)) MIN(sessions.MEAN(transactions.amount)) MEAN(sessions.SKEW(transactions.amount)) MEAN(sessions.NUM_UNIQUE(transactions.product_id)) MEAN(sessions.MAX(transactions.amount)) MEAN(sessions.SUM(transactions.amount)) MEAN(sessions.COUNT(transactions)) MEAN(sessions.MIN(transactions.amount)) MEAN(sessions.STD(transactions.amount)) MEAN(sessions.MEAN(transactions.amount)) NUM_UNIQUE(sessions.MODE(transactions.product_id)) NUM_UNIQUE(sessions.DAY(session_start)) NUM_UNIQUE(sessions.WEEKDAY(session_start)) NUM_UNIQUE(sessions.YEAR(session_start)) NUM_UNIQUE(sessions.MONTH(session_start)) MODE(sessions.MODE(transactions.product_id)) MODE(sessions.DAY(session_start)) MODE(sessions.WEEKDAY(session_start)) MODE(sessions.YEAR(session_start)) MODE(sessions.MONTH(session_start)) NUM_UNIQUE(transactions.sessions.device) NUM_UNIQUE(transactions.sessions.customer_id) MODE(transactions.sessions.device) MODE(transactions.sessions.customer_id) customer_id 1 60091 8 3 mobile 9025.62 40.442059 139.43 0.019698 5.81 71.631905 126 5 4 18 17 1994 2011 7 4 0 6 -0.476122 40 1057.97 78.59 312.745952 582.193117 0.589386 0.000000 7.322191 279.510713 4.062019 6.954507 13.759314 0.640252 5 1613.93 25 26.36 46.905665 88.755625 0.000000 -0.780493 0.778170 1.946018 2.440005 -0.312355 -0.424949 -1.038434 5 118.90 809.97 12 30.450261 50.623125 -0.059515 5.000000 132.246250 1128.202500 15.750000 9.823750 39.093244 72.774140 4 1 1 1 4 1 2 2014 1 3 1 mobile 1 2 13244 7 3 desktop 7200.28 37.705178 146.81 0.098259 8.73 77.422366 93 5 4 18 15 1986 2012 8 4 0 6 -0.277640 35 931.63 154.60 258.700528 548.905851 0.509798 0.000000 17.221593 251.609234 3.450328 15.874374 11.477071 0.755711 5 1320.64 18 56.46 47.935920 96.581000 0.000000 -1.539467 -0.440929 -0.303276 2.154929 0.013087 0.235296 -0.763603 5 100.04 634.84 8 27.839228 61.910000 -0.039663 5.000000 133.090000 1028.611429 13.285714 22.085714 36.957218 78.415122 4 1 1 1 1 3 1 2 2014 1 3 1 desktop 2 3 13244 6 3 desktop 6236.62 43.683296 149.15 0.418230 5.89 67.060430 93 5 1 21 13 2003 2011 11 8 4 5 2.286086 29 847.63 66.21 257.299895 405.237462 0.429374 0.408248 10.724241 219.021420 2.428992 5.424407 11.174282 0.854976 5 1477.97 18 20.06 50.110120 82.109444 -2.449490 -0.941078 2.246479 -1.507217 1.000771 -0.245703 0.678544 -0.289466 4 126.74 889.21 11 35.704680 55.579412 0.381014 4.833333 141.271667 1039.436667 15.500000 11.035000 42.883316 67.539577 4 1 1 1 1 1 2 2014 1 3 1 desktop 3 60091 8 3 mobile 8727.68 45.068765 149.95 -0.036348 5.73 80.070459 109 5 2 15 8 2006 2011 8 4 1 4 0.002764 37 1157.99 131.51 356.125829 649.657515 0.387884 0.517549 3.514421 235.992478 3.335416 16.960575 13.027258 0.382868 5 1351.46 18 54.83 54.293903 110.450000 -0.644061 0.027256 -0.391805 0.282488 2.103510 -1.065663 1.980948 -0.711744 4 139.20 771.68 10 29.026424 70.638182 0.000346 4.625000 144.748750 1090.960000 13.625000 16.438750 44.515729 81.207189 5 1 1 1 1 1 2 2014 1 3 1 mobile 4 5 60091 6 3 mobile 6349.66 44.095630 149.02 -0.025941 7.55 80.375443 79 5 5 28 17 1984 2010 7 7 5 5 0.014384 30 839.76 86.49 259.873954 472.231119 0.415426 0.000000 7.928001 402.775486 3.600926 4.961414 11.007471 0.602209 5 1700.67 18 20.65 51.149250 94.481667 0.000000 -0.333796 0.472342 -0.317685 -0.470410 0.204548 0.335175 -0.539060 5 128.51 543.18 8 36.734681 66.666667 0.002397 5.000000 139.960000 1058.276667 13.166667 14.415000 43.312326 78.705187 5 1 1 1 3 1 2 2014 1 3 1 mobile 5
I'm thinking, what kind of predictors can be invented with the regression channel?
I came to the conclusion that it is better not to think and not to invent, but to write a sort of re-processing algorithm that will synthesize the signs itself and check them, if the sign is good, then leave it if bad, then throw it away, so you can go through millions of options, it is clearly more efficient than human inventing.
Then you can improve and modify the good attributes, and then you can go through them again, and again, and so on, until the error drops.
I was inspired by Ivakhnenko's writings and the MGUA method. I like the philosophy of the method itself.
think or proven? what is interesting there?
I wrote that I have not been able to run the package, how can I prove anything if I have not even touched it, I read the documentation, there are chips, it makes sense to try, but I have not tried it for the reasons described above
you can't find the medium, std and mod by yourself.
Well, in general, it splits well... for any tai series is what you need. So you don't have to do it by hand.
This should be put in the LSTM, which were mentioned here
you can't find the medium, the std and the fashion yourself.
It's not all that primitive. Read more.
But in the end you will have to write your own feature synthesizer
it's not all that primitive, read more
i'm almost done with my bot! testing will be done soon
how many geomoro urges had to be overcomeI'm almost done with my bot! tests will be coming soon
wait)
He's been trading on neurons for three years now. I talked to him personally. He takes accounts of at least 100,000,000. Open his profile and you'll see all his accounts. He did it. So you can do it too. If you don't do it now, you'll do it later. Don't give up.)
He did it, so he has been trading on a demo account for three years.)
He succeeded and that's why he's been trading on a demo account for three years now ?)
there you go.
They trampled the illusion of enrichment...