User Tools

Site Tools


private:financial_prediction

Financial Prediction

Use DNNs to generate features from the time series and train a LMS predictor on top of them. This has the advantage that the features are static and the predictor can be adapted in an optimal Bayesian way.

Domains: bitcoin or forex - both have to be careful that changes are non-Gaussian.

What does “2 state HMM” mean?

private/financial_prediction.txt · Last modified: 2018/07/07 07:56 (external edit)