A forecast is intended to capture genuine patterns and relationships that exist in the historical data. Prediction of the next values of a time series before they are observed is difficult because of extrapolating beyond the limits of available data.
The aim of this project was to compare performance of parametric and non-parametric time series models, by means of simulation, in predicting the trends into the future.
Important part of the project was application built using R language at Shiny platform. My demo provides a graphical view over forecasting time series with parametric and non-parametric models.
To read more about the project, follow this link. To check out the app at Shiny platform, click here.
To download the summary in the form of report or poster, click this link.
*Please be patient while waiting for the result. Computational method INLA and its interface R-INLA are powerful but it may take few more seconds! Obviously, the speed of your broadband connection matters as well.
**The application was tested using Firefox and Google Chrome browsers only. It will not work loaded in IE as it does not correctly render IFRAME.
If you like what you’ve just read, subscribe to regular email updates sent right into your inbox. Hanging out elsewhere? Say hi from anywhere, anytime ->