Softmax Regression and Classification of future returns

[kingdezee]

Can anyone code a Softmax Regression, using these variables - 1 day return, 5 days return, ratio of intraday high and low, changes in daily high, low, open and close prices in two consecutive days, difference between morning and afternoon volume and their daily changes, applying the standard deviations across each variable? and perhaps pit the variances between two correlated instruments such as the S&P500, and the VIX index?

 Can this also be done for other varied instruments such as major indexes which are have some relation to each other? 

[Stanislav Korotky]

Could you probably try the way described in this article?