I work with a company who model financial markets using machine learning models. Currently one of our models uses TA-lib as it's 'features' (model inputs) I want to to start writing custom indicators for this model in python. Does anyone one know where to start, any online guides ?
Thanks
EDIT: I trade on a daily basis and understand the mathematics behind these indicators. I'm not the greatest programmer but this task is certainly within my reach. Essentially our model takes in 150 indicators and finds correlations between the outputs of these and price. The task now is to create custom, domain specific indicators which help the model make better predictions. I should mention we're modelling alt coins (litecoin, dash, DOGE).
One problem I can see is that the predictions are based only on data from each crypto coin, they have no outside influence. The indicators I would like to design would contain data regarding Etherum and bitcoin. It seems alts are correlated to ETHs price and conversely anti correlated to BTCs price. It think it would be interesting to design an indicator which shows the price change, the idea being moves in BTC and ETH have an effect on the alt coin prices. Then use the ML to see if it can detect any patterns.
Additionally, i should mention we're looking for new partners with experience
- Trading systems / algorithmic trading strategies
- Feature engineering and cryptocurrency domain knowledge
Submitted April 06, 2016 at 09:05AM by loba333 http://ift.tt/1Wcg9Qs
No comments:
Post a Comment