mimtek Posted December 9, 2017 Report Share Posted December 9, 2017 Machine learning is the best choice to develop based strategies ? Quote Link to comment Share on other sites More sharing options...
mike83 Posted December 9, 2017 Report Share Posted December 9, 2017 machine learning , AI and data science wont work for stock market trading especially intraday. may be it will be helpful for short term investments. that s why amazon ceo jeff left that financial software company. Also ML AI wont work efficiently in more than one core processors this is one reason neuro ml softwares worked well in xp and after that they failed with their prediction models for intraday or shorterm trading with dual core processors. We need GPU TPU processors to be cheap and until then we retail traders have to wait. Quote Link to comment Share on other sites More sharing options...
Albnd Posted December 9, 2017 Report Share Posted December 9, 2017 machine learning , AI and data science wont work for stock market trading especially intraday. may be it will be helpful for short term investments. that s why amazon ceo jeff left that financial software company. Also ML AI wont work efficiently in more than one core processors this is one reason neuro ml softwares worked well in xp and after that they failed with their prediction models for intraday or shorterm trading with dual core processors. We need GPU TPU processors to be cheap and until then we retail traders have to wait. Completely wrong. I'm successfully using R with parallel computing (40-core CPU power spread in 4 VPS) and works like a charm. Machine learning requires a sound statistical background and is not as easy as go-up-go-down indicators. Quote Link to comment Share on other sites More sharing options...
Sixer Posted December 9, 2017 Report Share Posted December 9, 2017 Which program(s) are you using for your intraday predictions ? https://www.sendspace.com/file/y7wh74 Sixer laser1000it and mimtek 2 Quote Link to comment Share on other sites More sharing options...
laser1000it Posted December 9, 2017 Report Share Posted December 9, 2017 Which program(s) are you using for your intraday predictions ? https://www.sendspace.com/file/y7wh74 Sixer Che è sta roba? what's that ? Quote Link to comment Share on other sites More sharing options...
Sixer Posted December 9, 2017 Report Share Posted December 9, 2017 These are intraday chart projections and the proof that they are correctly predicted. Sixer Quote Link to comment Share on other sites More sharing options...
mike83 Posted December 9, 2017 Report Share Posted December 9, 2017 Completely wrong. I'm successfully using R with parallel computing (40-core CPU power spread in 4 VPS) and works like a charm. Machine learning requires a sound statistical background and is not as easy as go-up-go-down indicators. i just mentioned for neural networks predictions. rest all which comes with python R doesn't need TPU Quote Link to comment Share on other sites More sharing options...
mike83 Posted December 9, 2017 Report Share Posted December 9, 2017 (edited) These are intraday chart projections and the proof that they are correctly predicted. Sixer what software this is and what is the model and feature used? looks like pics were taken in 2010 is these signals generated in realtime market? many times model lookfwd and generate new one during back testing and wasted lots of my time with ml things Edited December 9, 2017 by mike83 Quote Link to comment Share on other sites More sharing options...
mimtek Posted December 9, 2017 Author Report Share Posted December 9, 2017 (edited) How do I use neuroshell vs order application can it be combined the most extensive c ++ library or softwore Support Vector Machines. Gradient Boosted Trees. Random Forests. Extremely Randomised Trees. Multi-layer Perceptron, aka Neural Network. Ensembles: Combine the forecasts of any number of predictors. Constant retraining, always adapting to the market. Filtering.DSP garnozel filter vs vs _? input - predictions - output for example extern "C" __ declspec (dllexport) void _cdecl predictions (double * input, double * output,long ınt sıze) ............. ............ return: Personal experience I work with Neuroshell and PolyAnalyst actuel 1 lead norm input norm 1lag (input norm) 2lag (input norm) 3lag (input norm) 4lag (input norm) output 1 lead norm SVM REG the output is super - the real trade sucks Edited December 9, 2017 by mimtek ESVepara 1 Quote Link to comment Share on other sites More sharing options...
Uiyutr Posted December 9, 2017 Report Share Posted December 9, 2017 Which program(s) are you using for your intraday predictions ? https://www.sendspace.com/file/y7wh74 Sixer Does it use of some kind of ARIMA? or its something else like neural Network? Can you please explain a little more? Quote Link to comment Share on other sites More sharing options...
mike83 Posted December 10, 2017 Report Share Posted December 10, 2017 Does it use of some kind of ARIMA? or its something else like neural Network? Can you please explain a little more? all those pics were taken in 2010. not sure whether model can still work for 2017 data. that's said all. Quote Link to comment Share on other sites More sharing options...
Sixer Posted December 10, 2017 Report Share Posted December 10, 2017 Did anybody test the accuracy of the following indicators with Metatrader platform ? Metatrader 5: AR extrapolation of price https://www.mql5.com/en/code/129 Metatrader 4: extrapolator https://www.mql5.com/en/code/8608 Sixer mimtek 1 Quote Link to comment Share on other sites More sharing options...
mimtek Posted December 10, 2017 Author Report Share Posted December 10, 2017 (edited) Did anybody test the accuracy of the following indicators with Metatrader platform ? Metatrader 5: AR extrapolation of price https://www.mql5.com/en/code/129 Metatrader 4: extrapolator https://www.mql5.com/en/code/8608 Sixer If I have c/c++ code, I can try pls hlp http://www.metatraderfiles.com/fourier-extrapolator/ Edited December 10, 2017 by mimtek Quote Link to comment Share on other sites More sharing options...
mj10 Posted December 23, 2017 Report Share Posted December 23, 2017 i have tried many software with different inputs the results were mix not trade able. The problem is prices are not fix they go up and down if there is way we can make price straight then machine learning can be used effectively. Quote Link to comment Share on other sites More sharing options...
mike83 Posted December 25, 2017 Report Share Posted December 25, 2017 stock market is based on brownian motion can't be applied or predicted with ml softwares. may be once u can use them for stastics and research works or else in low volume or operator driven scrips not for cme nyse nikkei exchanges Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.