SCM 303 Lecture Notes - Lecture 8: Simple Linear Regression, Linear Regression, Mean Absolute Percentage Error
15 Oct 2016
School
Department
Course
Professor
Document Summary
Forecasts predict the future (demand levels, supply levels, and prices: usually demand volumes (downstream) Example: how many iphone 6s plus phones will apple sell next month, next quarter, or next year: also used for supply volumes & supply prices (upstream) Example: how many microchip factories will operate next year and what. Critical information for planning is their combined output: production capacity, distribution capacity, revenues, supplier contracts, new products. The future is always uncertain: forecasts are almost always wrong, forecasts tend to be more accurate for near-term periods, more accurate for short term than long term, groups of products vs. Individual items: better forecast for similar products, more accurate forecast when lumping a bunch. E. g. dark green cars versus all cars: forecasts are no substitute for calculated values, always better to calculate. Forecast error: the difference between the actual demand the forecasted demand for a given time period (difference between dt and ft)
