Forecasting Flashcards
What is a causal forecasting method
Forecasting based on the assumption that certain variables have a causal effect on the future f.ex regression analysis
What are qualitative forecasting methods
Like everything qualitative in academia it involves talking with experts
Is time series forecasting qualitative or qunatitaiteve
Quantitative like causal forecasting
What is the seasonal component within a time series
Any regularly repeating pattern with a cycle smaller than one year
What is the cyclical component within a time series
Any regularly repeating pattern with a cycle loner than one year
What is the irregular component within a time series
The unknown unaticipated patterns. It is the residual besides the cyclical and seasonal components
What are smothing methods
Ways to filter out the irregular compponent from a time series
What is the moving average
The average for a recent period of time, as time flows the moving average tags along, abandoning the oldest values for the latest
How is forecast accuracy most offten mesured
By the average sum of squared errors MSE (mean squared errors)
For weighted moving averages the sum of the weights is 5
False it is 1
Exponsntial smothing provides a forecast with cis a weighted average
Yes, all forecasts can be interpreted as weighted averages
Why is a small smothing constant perferred for exponsntial smothing in a timeseries with substantial random variability
Becouse we do not wnat to overcorrect for variability. However it the variability is not random a larger constant will allow for faster correction
How should you choose the smothing constant and the number of periods to include in moving average
The number that minimizes the mean squared error
Explain the exponential smothing formula
F(t+1) = aY(t) + (1 - a)F(t) where F is forecast, Y is actual value and a is the smothing constant and t indicates time period
What is the difference between simple and multiple regression analysis
The number of explanatory variables
What are auto regressive models
Regression models where the independent variables are previous values in the timeseries
What is the delphi method of forecasting
Each expert tells their idea listens to others ideas and than reconsiders based on what theyve heard until somewhat of a consensus is reached
