I’ve been working recently on a project in the e-business space to forecast customer demand. The data consist of multiple time series comprised of dates and usage numbers. From the existing data, the challenge is to extrapolate into the future, making predictions as far out as 90 days or more. As always, I look to the R Project for Statistical Computing for help. And, as always, R delivers.
One thing I’ve found in my travels is that challenges surrounding forecasting in business often take precedence over statistical purity. Back in the day, I could devote special attention to each of a small number of forecasts I was responsible for, examining countless graphs, statistics, autocorrelation functions, etc. to choose a “best” model. That approach breaks down, however, when there are scores or even hundreds of separate series to forecast – as is often the case today. I now obsess on prediction quality over statistical purity, and generally prefer techniques that choose “optimal” solutions automatically based on performance criteria. I guess I’ve become a hardened algorithmist.
Register or login for access to this item and much more
All Information Management content is archived after seven days.
Community members receive:
- All recent and archived articles
- Conference offers and updates
- A full menu of enewsletter options
- Web seminars, white papers, ebooks
Already have an account? Log In
Don't have an account? Register for Free Unlimited Access