tag:blogger.com,1999:blog-21831384.post6035918188775576425..comments2019-10-03T14:11:35.530+05:30Comments on BzST | Business Analytics, Statistics, Teaching: "Predict" or "Forecast"?Galit Shmuelihttp://www.blogger.com/profile/06119270323184007583noreply@blogger.comBlogger4125tag:blogger.com,1999:blog-21831384.post-83656199618633306982012-01-16T11:31:51.318+05:302012-01-16T11:31:51.318+05:30Dushyant - I would just add that "prediction&...Dushyant - I would just add that "prediction" need not be only of the future. You could try to predict the response of some records (not necessarily future ones) based on responses of other records. For instance, you can try to predict the chance of fraud in some existing transactions, given a sample of transactions with known fraud/non-fraud labels.Galit Shmuelihttps://www.blogger.com/profile/06119270323184007583noreply@blogger.comtag:blogger.com,1999:blog-21831384.post-20504363663991356532012-01-16T11:18:58.494+05:302012-01-16T11:18:58.494+05:30Really nice discussion :)
Having worked in the An...Really nice discussion :)<br /><br />Having worked in the Analytics Industry for 5 years, I come more of the opinion that Forecasting and Prediction are completely different terms though both have been used interchangeably. <br /><br />Forecasting- A method of using historical data assuming that there is a trend to be followed and forecasting is basically the "prediction" of the trend!<Dushyanthttps://www.blogger.com/profile/00180094709837075512noreply@blogger.comtag:blogger.com,1999:blog-21831384.post-49779017962653810872011-09-06T15:18:43.634+05:302011-09-06T15:18:43.634+05:30Good point Rob. Another point: I find that the dis...Good point Rob. Another point: I find that the distinction between "forecast" and "predict" is more prominent with statisticians and econometricians (who often concentrate more on the model "appropriateness" than on predictive accuracy), to emphasize the difference between models that account for temporal dependence directly (such as ARIMA) and those that don't. Galit Shmuelihttps://www.blogger.com/profile/06119270323184007583noreply@blogger.comtag:blogger.com,1999:blog-21831384.post-74846762516656992872011-09-06T15:06:57.494+05:302011-09-06T15:06:57.494+05:30I've heard the time series/cross sectional dis...I've heard the time series/cross sectional distinction before, although I tend to use predict and forecast interchangeably in both contexts myself. But the aggregate/individual distinction is new to me and seems impossible to use consistently in practice. How much disaggregation is required before you would switch from forecasting to prediction?Rob J Hyndmanhttps://www.blogger.com/profile/14639396933879794459noreply@blogger.com