Machine Learning

To leverage the advancements in machine learning and more, we've added a new module to calculate the Linear Regression via machine learning. The new app 'Linear Regression'. The app will use the machine learning i.e.

  • The user will be able to select any global instrument e.g. Stocks, ETFS, Funds, Index, Commodities, Currency, Crypto

  • Set the duration of dates

  • The app will then take daily historical returns for the said symbol

  • Calculate the daily LAG1 returns i.e. Return of 1-Day before for each day

  • Run machine learning model to predict returns on above data set

  • Calculate the linear regression model by apply the fit for

    • Daily LAG1 return

    • Daily return

  • Calculate the linear regression line

  • Calculate the predicted returns based on the linear model

  • Provide a scatter plot to visualize the daily LAG1 returns, returns and regression line

The time series chart also provide a way to view the returns vs predicted returns and to see how many times the predicted returns direction was correct i.e. If the predicted returns and daily returns directions are same, it is considered a good forecasted direction.

Also it provides stats like;

  • # number of observations i.e Number of trading days,

  • # of positive forecast

  • % of positive forecasted returns within the model

This is the very first Machine Learning model among the many more models to come at

To access: From the top bar, Click ML > Linear Regression

Last updated