The object of this project is getting a better understanding of price movements in crypto markets: can you make (some) money using autoregressive neural networks?

Installation (local)

$ git clone your_project_name


Copy data to data/raw and run scripts in appropriate order

> source("01_preprocess.R")
> source("02_train.R")
> source("03_predict.R")


If you?re using RStudio, press Cmd/Ctrl + Shift + T (or run devtools::test() if not) to run all the tests in a package

How it works

Basically, the trained model performs two steps: 1. for each crypto market, predict the next hour price, using an autoregressive neural network 2. improve predictive power for each market by combining these predictions

Oxylo/canaries documentation built on Oct. 30, 2019, 10:23 p.m.