Below is the first entry in a likely long enduring series of security specific forecasts that I will be generating for the Dow constituents and ultimately for all S&P 500 candidates. The underlying forecasts were generated using a proprietary series of algorithms built in R that, through a series of cross validation tests, select an optimal array of factors that feed into a number of final models (multiple regression, neural network, logistic regression, etc). 7 cross validation 'folds' were used and a final hold out sample was utilized to confirm and backtest results. In future posts I will publish the results of the backtest model and provide context on the top 5 factors that feed into each forecast model. Each forecast is generally based on 15 indicators that all vary from security to security.
As you can see, some Dow constituents are projected to increase over the next year though there are a few outliers. Specifically, energy company share prices should improve as market sentiment shifts and banking stocks the same as investors react to the new interest rate/money supply paradigm.
Simply click the image below to see specific detail on the constituent forecasts. If you would like the underlying data feel free to message me or run the image through an OCR converter...
No comments:
Post a Comment