Week 11 - Forecasting 3

Question 1

What are causal models used for? (2)

Answer 1

• To identify variables, or a combination of variables, which affect demand and ar ten used to predict future levels of demand
• Used in linear regression and multiple regresssion

Question 2

What is regression analysis?

Answer 2

Is the development of a mathematical equation that predicts the value of a dependent variable from simple regression or multiple regression independent variables

Question 3

Regression analysis example

Answer 3

Sales (dependant) is related to advertising spend (independent)

Question 4

What is the difference between regression analysis and causal forecasting?

Answer 4

Causal forecasting relates the forecasted variable to variables that are supposed to influence or explain that variable

Question 5

What is the benefit of having a model that relates inputs to outputs?

Answer 5

It facilitates a better understanding of the situation and allows experimentation with different combinations of inputs to study their effects on the forecast (output)

Question 6

Which type of graph is best suited for linear regression

Answer 6

Scatter graphs

Question 7

Characteristics of a linear regression scatter graph (3)

Answer 7

• A small scatter indicates a strong relationship
• A positive relationship occurs when one variable get larger the other variable also gets larger
• A negative relationship occurs when one variable gets larger the other gets smaller

Question 8

Linear regression straight line equation

Answer 8

Question 9

How to answer a linear regression question?

Answer 9

Question 10

Interpreting linear regression results (2)

Answer 10

• The coefficient of determination (R2) shows the proportion of the variation in Y that is explained by X
• The coefficients of the regression equation are given as intercept (b0) and the X variable 1 (b1)

Question 11

3 sections of regression analysis

Answer 11

• Summary output
• Anova
• Unlabbled section

Question 12

Summary output in regression analysis (4)

Answer 12

• Multiple R is the sample correlation coefficient
• R Square is the coefficient of determination (R2) - this shows the proportion of the variation in Y that is explained by X
• Adjusted R square modifies R2 by incorporating the sample size and the number of explanatory variables
• Standard error - is the variability of the observed Y values from the predicted values (the amount of scatter)

Question 13

ANOVA regression analysis

Answer 13

• Concerned with the value of the significance of F
• If the significance F is less than the level of significance typically (0.05) we would reject the null hypothesis
• If we reject the null hypothesis, then we may conclude that the slope of the independent value is not zero and therefore, is statistically significant in the sense that it explains some of the variation of the dependant variable around the mean