Predict bitcoin price using machine learning tools based on the data from bitcoin_trade_data project
After finished this project, will start another project to autotrade the future in OKex based on machine learning result.
##Updateds
Adding Analytic scripts(to be competed)
finished data pre-processing
bfx_bids_wall: bids wall from bitfinex (select the largest bids amount from order book)
bfx_asks_wall: asks wall from bitfinex (select the largest asks amount from order book)
bfx_total_bids: total bids from order book(bitfinex)
bfx_total_asks: total asks from order book(bitfinex)
bfx_buy_volumn: total buy volumn from bitfinex during this period
bfx_sell_volumn: total sell volumn from bitfinex during this period
USDT_exceed: Huobi OTC price - USD/CNY enchange rate (This index is to see if the Chinese prople are willing to buy bitcoin with a premium price.
pre_price15: bitcoin future quarterly price from okex 45 minutes ago.
pre_price10: bitcoin future quarterly price from okex 30 minutes ago.
pre_price5: bitcoin future quarterly price from okex 15 minutes ago.
pre_bfx: bitcoin price from bitfinex 3 minutes ago.
pre_news10: Get news from wallstreet.cn and analyze the emotion(from 0-1). Pre_news10 is the emotion 30 minutes ago
next_price15: bitcoin future quarterly price from okex 45 minutes later.
next_price10: bitcoin future quarterly price from okex 30 minutes later
next_price5: bitcoin future quarterly price from okex 15 minutes later
Data Modeling: Use sklearn - Test SVM, Random Forest and neural network with different parameters. Try to find out the best strategy.(To be updated)
1 Bug fixed