kartikmadan11/metatraderforecast

RNN based Forecasting App for Meta Trader and similar trading platforms

Hi! It's me, Kartik Madan

👨‍💻  About Me:

I am a Full-Stack Developer from New Delhi, India.

• 🚀 I’m working as a Specialist Programmer at Infosys and building the Infosys Equinox platform
• 🔭 I work primarily on the MEAN stack and Microservices. I also have a deep interest in Deep Learning and Recurrent Neural Networks
• 🖋️ Follow me on Medium
• ⚽ I love playing and watching football

Find me on

---

🛠  Languages and Tools :

                         

metatraderforecast's Issues

Do you have any time to show how to handle more than a close price as input in python?

Hi,

I'm struggling to add some other parameters in the python code...such as MACD, RSI for example along with the Close. Quick example in python would be appreciated.

Kind Regards

MultiCharts does not support sockets

Using Multicharts .Net 64 C# based indicator as a socket client for socketserver.py throws error.

please help me

I'm not familiar with these. Can you write down how to use your code，thinks

dude, this code is looks like very interesting

can you record a video, I didn't understand how it works, and solving which problems in which ways?

Make it Multi-Feature

Hello,
I would like to make it multi-feature, i mean for now it is using sole feature (closing price), i have more indicators values, thos values have a values similar to prices format. but they are in csv file not from the MT4.
how can i involve them in the training?
BTW< the csv file is updating every hour also, it is using another python script to keep it updated

Poor documentation, please help

How is this supposed to work? After adding the Json library to the folder in MQL5, I was able to compile the EA, but once added to the chart and pressed the "train" button, nothing happens. I have made sure to leave the socketserver running.
I am not using a GPU, the socketserver is running on anaconda env, is it a variable PATH problem?

AIMetatrader

The question is how you can use the publication as an open source or publicly accessible, because GitHub and Tensorflow are legally provided.
Ultimately, I now have a hobby website where I use Metatrader as an inframe and selected this AI. I call the project AIMetatrader and see the whole thing as a hobby side. I had no breakthrough with my meager knowledge. I haven't been able to install the AI yet. Can i hire you

build.save_plot() does not work properly

The exported test vs prediction plot gives unexpected/erroneous results. Like the following exported image shows.

This does not happen however when using python directly without sockets. Following image is running on similar data.

Perhaps using Plotly could solve above problem

error in serv.recvmsg()

``
D:_Python\python.exe "C:\Users\Beertje\PyCharm Community Edition 2019.2.4\helpers\pydev\pydevd.py" --multiproc --qt-support=auto --client 127.0.0.1 --port 52744 --file "D:/_Forex Projects/python/MetaTraderForecast/socketserver.py"
Socket Created at ('127.0.0.1', 9090). Waiting for client..
connected to ('127.0.0.1', 52997)
Traceback (most recent call last):
File "C:\Users\Beertje\PyCharm Community Edition 2019.2.4\helpers\pydev\pydevd.py", line 20, in
from _pydevd_bundle.pydevd_constants import IS_JYTH_LESS25, IS_PYCHARM, get_thread_id, get_current_thread_id,
File "C:\Users\Beertje\PyCharm Community Edition 2019.2.4\helpers\pydev_pydevd_bundle\pydevd_constants.py", line 192, in
from _pydev_imps._pydev_saved_modules import thread
File "C:\Users\Beertje\PyCharm Community Edition 2019.2.4\helpers\pydev_pydev_imps_pydev_saved_modules.py", line 22, in
import xmlrpc.server as _pydev_SimpleXMLRPCServer
File "D:_Python\lib\xmlrpc\server.py", line 108, in
from http.server import BaseHTTPRequestHandler
File "D:_Python\lib\http\server.py", line 102, in
import socketserver
File "D:_Forex Projects\python\MetaTraderForecast\socketserver.py", line 71, in
msg = serv.recvmsg()
File "D:_Forex Projects\python\MetaTraderForecast\socketserver.py", line 60, in recvmsg
self.conn.send(bytes(train_test_model(self.cummdata), "utf-8"))
File "D:_Forex Projects\python\MetaTraderForecast\socketserver.py", line 24, in train_test_model
train = build.train(training_set=data[:trainSize], date=date[:trainSize], lr=input_data['LearningRate'], scale=input_data['Scale'], epochs=input_data['Epochs'], momentum=input_data['Momentum'], optimizer=input_data['Optimizer'], file_name=file_name, architecture=input_data['Architecture'])
NameError: name 'build' is not defined
<class 'dict'>

Process finished with exit code 1

``
I get the following errors while running. Thanks for your help!

Update of build.py for TensorFlow 2

Hey everyone,

I was testing this out and found a few changes that need to be made to the build.py script.

Remove:

1. CuDNNLSTM
2. CuDNNGRU (they are no longer called like that).

Substitute:

1. tf.ConfigProto with tf.compat.v1.ConfigProto
2. tf.Session() with tf.compat.v1.Session
   Same parameters, just change the syntax.

This is just a quick fix and should be rewritten for the new version of tensorflow.

Great work kartikmadan!

Is it possible to use this in MT4 Tester?

Is it possible to back test using this system via the MT4 Tester or would the barrage of data stop it being able to make trades in the necessary time frame?