MGMT 333 Lecture Notes - Lecture 2: Moving Average, Forecast Error, Average Absolute Deviation
11 Oct 2016
School
Department
Course
Professor
Document Summary
We (cid:271)reak the pla(cid:374) do(cid:449)(cid:374) i(cid:374)to sku"s, the(cid:374) (cid:373)ore detailed fore(cid:272)ast. Qualitative forecasting lets us look at new products and technology. Random variation and trends is what we care about. Pick model based on: horizon to forecast, data availability, accuracy required, budget, available personal. Basic way to calculator a forecast is to take a simple moving average. The benefit of taking a simple moving average is it takes out random variation. 3 week has variation very reactive: less inputs so they are worth more; 33% fashion industry would like this, quick changing. 6 week is 16% so less reactive steady product that is mature. Fore(cid:272)ast (cid:271)ias: it"s a tre(cid:374)d i(cid:374) this (cid:272)ase. Trends in forecast make it lag and to constantly be biased in a single direction. It is a negative consequence, if there is any trend. If you ha(cid:448)e tre(cid:374)d do(cid:374)"t use a (cid:373)o(cid:448)i(cid:374)g a(cid:448)erage.
