BUAD 310 Case Analysis
(2) Regression Analysis
1. According to the graphs that we built for the variables, we found the distribution of the variables (opening, budget) to be abnormal. The opening distribution has two outliers, and is right skewed. And the budget distribution is right skewed as well. These two variables might have contributed to the outliers and the right-skewed distribution of the response variable- US Revenue. Yes, this should worry us for the outliers will pull the regression line towards itself, distorting the regression result.
2.
Correlation matrix:.
According to the pair plot and correlation matrix we built for the variables, US Revenue seems linearly dependent on Budget and Opening and not so much on Opinion. However, transformation between US Revenue and Theaters might be needed. Nothing unusual stands out.
3.
Multiple linear regression results:
Dependent Variable: US Revenue
Independent Variable(s): Budget, Opening, Theaters, Opinion
Parameter estimates:
Analysis of variance table for multiple regression model: Summary of fit:
Root MSE: 15.69288
R-squared: 0.9807
R-squared (adjusted): 0.9782
Fitted Regression Line: ŷ= -67.724+ 0.135(Budget) + 3.017(Opening) -0.002(Theaters) + 10.262(Opinion)
R^2=0.9807: About 98.07% of the variation in the US Revenue values is accounted for the regression onto the predictor variables Budget, Opening, Theaters, Opinion.
Interpretation:
b0 = -67.724:
When Budget= Opening=...