Historical Data Management and Configuration for a Neural Net Fitting Process

GUI for Testing Connection, Save and Manage Datasheets

TWS API for a simple request and managing of historical Stock Information
Explore the docs »

Request Feature · Report Bug

About The Project

Neural Net Fitting of given historical Market Data. The Given Arguments can be checked for Input and Output values. After Setting Up Input & Output, you be able to create a neural Net (see https://de.mathworks.com/help/deeplearning/gs/fit-data-with-a-neural-network.html)

 nftool

v.1.0 receives historical information. the command (see https://de.mathworks.com/help/releases/R2020b/trading/ibtws.history.html)

 d= history(ib,ibContract,startdate,enddate,ticktype,period,tradehours)

includes:

• Numeric representation of a date

• Open price

• High price

• Low price

• Close price

• Volume

• Bar count

• Weighted average price

• Flag indicating if there are gaps in the bar

Getting Started

Make sure you have the latest stable version of TWS API see: https://interactivebrokers.github.io

If you have a MATLAB license, just get all .m and the .fig files and Run

 Programm.m

If you dont have MATLAB, install the Windows Excecutable TWS_API_Historical-Data-Manager.exe

If you are able to Run the GUI, make sure youre Trader Workstation is set up correctly with input arguments you choose in the GUI

How it works

• Press 'Connect' to run the given arguments from the API-Config TWS Panel and the Input Panel.(see RealtimeStreaming.m)

  • Make sure the Stock is listed in your watchlist in IB Trader Workstation (TWS)

• Press 'Receive Data' button to get the historical Data from your liveStream (see Plotting.m)

• Select Input and Output arguments to build your own neural Network with nftool

• Press the 'Neural Net Fitting' button (see neuralNetFitting.m)

• Save the historical Data by clicking 'Save to .xls'.

• The Sheet will be located where the Programm.m or .exe is located

• Press 'Disconnect'

• Start your next analysis