click below
click below
Normal Size Small Size show me how
PROD.MGMT 2
Quantitative Forecasting (Time Series)
Question | Answer |
---|---|
Define Casual Forecasting (regression analysis). | Assumes that the factor to be forecasted exhibits a cause-effect relationship with one or more independent variables. |
What is the objective of Casual Forecasting (regression analysis)? | To discover the form of that relationship between the variables and be able to forecast future values of the dependent variable. |
What method is used in Casual Forecasting? | Simple Linear Regression model (SLR) |
What is the Simple Linear Regression (SLR) model? | Finding the best fitted linear line for a given set of data. |
What is the formula for Simple Linear Regression? | y=a+b*x y: predicted(dependent variable) X: predictor (independent) b: regression slope of line a: Y intercept |
How to find the slope or B in Simple Linear regression? | y=n(sum of xy)- (Sum X)(Sum Y) ------------------------ n(sum x ^)- (sum x ^) |
How to find Y intercept or A in simple Linear Regression? | a= (sum Y)- b(sum x) ---------------- n |
What is Time Series Analysis? (TS) | Set of measurements ordered through time on a particular quantity of interest |
Components of TS? | T: Trend S: Seasonality C: Cyclical R: residual |
What are smoothing methods? | methods used in adjusting data to cancel out the effect of random variations and reveals components that we are looking for. |
What are the popular smoothing methods? | 1. Moving Average (MA) 2. Simple Exponential Smoothing (SES) 3. Naive Approach |
What is the Naive Approach? | Using data from previous period to forecast future values |
What is the Moving Average (MA)? | Based on idea that any large random component at any point in time will have smaller effect if the value is averaged with neighbors. |
When does Moving Average (MA) not work? | When trends or Seasonality occurs |
Formula for MA? | MA= Sum (most recent N data values) ------------------------- N |
What does having a large N for MA mean? | Slow response to real changes in demand, more stable (greater smoothing effect) |
What does a small N for MA mean? | Quick response to changes in demand, little smoothing effect. |
What is weighted Moving average? (WMA) | same as MA except putting more weight on recent data |
What is Simple Exponential Smoothing? (SES) | smoothing method giving weight to all data but it gives weight to recent observations more heavily |
When does Simple exponential smoothing (SES) work well? | when there are no seasonal or trend components |
what is the formula for SES? | f(t)= a* A t-1 + (1-a)Ft-1 |
When should one use a small (a) in a TS? | If the TS contains substantial random components don't give much weight to most current observation alone. |
When should one use a Large (a)in a TS? | If TS is rather smooth |
What can a higher value of (a) do? | respond to sudden change more quickly |
What is Linear Trend Approach? (time series regression forecasting) | Deals with trend component. Based on Least square approach. |
Which forecasting method provides most accurate forecast? | Linear Trend Approach |
What is diagnostic Checking? | Evaluating performance of a regression equation |
what are four forms of diagnostic checking? | 1. Residual analysis 2. MSE 3. T test 4.R^2 |
What is MSE? | Mean squared Errors. Smaller the better. Useful when two or more regression models are considered. |
What is the t- test? | Test for significance (usefulness) of regression equation. Testing weather I variable effects D variable. |
What is Correlation Analysis? | Measure of linear association between x and y. determines relative strenght or relationship |