Hey guys, welcome to Redora, this is a trading prediction bot comprised of various types of Neural networks to make accurate predictions for your Financial timeseries. I initiated this project initially to predict Bitcoin Stocks, however due to the volitilty of cryptocurrency, i shall just leave it to trading stocks. Be aware you may need to have a powerfull computer to run it ! also do not tamper too much with the sequence length and epoch i almost fried my processor but putting epoch at a high value as well as sequence length.
Ok so youll need alooot of dependencies !!! i used many libraries in my program because due to the complexitity of enural networks i would not have been able to complete such a ambicious project with straight hard coded functions. If you plan to further develop Redora you may need the following dependancies installed:
Matplotlib
Random
numpy
csv
lstm
time
tensorflow
sklearn.svm import SVR
pandas
sklearn.preprocessing import MinMaxScaller
tkner
This project was coded in Python so you will need to know about the ins and outs of python prior to working on Redora, not much just a fresher course to recap your programming skills.
If you wish to run the whole program in a comiled version, go to the comiled file and run the file named:
graphing.py
This version is a development version of Redora so ull need to have python installed on your PC.
For anyone wishing to help me develop my project further feel free to submit a pull request.
- Daniel Goncalves - Initial work (2018)
See also the list of refrences in the research paper attached as all the tutorial links and blog articles really helped alot in development.
This project is licensed under the MIT License and is an open source development project to be used mainly for research of algorithmic trades.
- Many thanks to Saraj Raval a youtuber who helped me understand Neural Networks throughly