MGT 3121 Lecture Notes - Lecture 3: Time Series, Exponential Smoothing
11 Feb 2017
School
Department
Course
Professor
Document Summary
Cheat sheet contains formulas, examples, definitions, however limited to one page. Moving part: whenever you have new data, you always keep n periods data in calculation. Different weightings are assigned to demand values. More recent data should be weighted heavier. Every step of way we correct for previous mistake alpha= how aggressive you want to correct for it. Mean square is the highest of all the three (for es) Why: need to have a large sample size. Your inferences, confidence level interval will be reliable. When you have more than 30 ( consider you have a large data set) Don"t bet your life on 3 observations. Drawbacks with averaging-> it is hard to see a trend. Very hard to capture what"s the trend. Underlying assumption of time series forecasting: what has happened before will continue to happen in the near future.
