My R code for Forecasting
These are my efforts for creating R code that forecasts sales or financial instruments.
1)R Folder: Statistical models either in plain or function form apply techniques from the forecast package and prophet from Facebook. The example is on monthly items sales.
There is a first stage, where the techniques are trained with the 75% of the sample and then tested against the remaining hold-out sample. Then based on RMSE we select a winner, which in this case is prophet and train it with the whole sample so as to predict the next month's sales.
2)Root: The financial indices code applies the same techniques on the % difference of closing from opening price of those indices It seems that prophet performs better. With this piece of code you can do a historical prediction and measure the accuracy of each technique for each index. You can also add your own symbols(shares etc).
3)Shiny Folder:The Shiny app with prophet is a demo app for the prophet algorithm by facebook. User can upload a csv file containing a column with dates and a column with the time-series.