Dependency techniques, regression analysis Flashcards
When do you use the adjusted R-squared?
when you are compare different models, otherwise normal r-squared
If u keep adding more and more independent variables in an OLS regression, what happens with r-squared?
if u keep adding more and more independent variables, r-squared would go up
If the significance level is below the classical standard cutoff value 0.05 what does it mean?
the statistic is significant. r-squared is significantly different from 0; its a good regression model!
What does the beta coefficient show?
the relative affect of the variables on the dependent variable
What is multicollinerity? (an assumption for regression)
multicollineraity: between the independent variables. best case scenario is when there is no correlation between independent variables because as the correlation rises between the independent variable you can reach a point where u cannot separate the variance between the variables. the problem of this is called multicollinearity
What is the standard recommended value to show whether there is multicoll or not? (VIF)
values at 10 or below
below 10 is ok
above 10 you have multicoll
What is the independence of error terms?
an assumption in regression
What is normally distributed residuals?
error terms. is an assumption in regression. Residuals are the differences between the observed and predicted values in a regression model. When these residuals are normally distributed, they form a symmetric, bell-shaped curve when plotted. This means most errors are small and close to zero, with fewer large deviations.
Normal residuals ensure reliable hypothesis testing and confidence intervals. If the residuals deviate from normality, it may suggest issues with the model, such as omitted variables, incorrect functional form, or heteroskedasticity, which might require further investigation or adjustments.
What is the linearity of the coefficient?
are the independent variables lineraly related to the dependent variables? ask for partial regression lots and you’ll see. refers to an assumption in regression analysis that the relationship between the independent variables (𝑋) and the dependent variable (𝑌) is linear in terms of the model’s coefficients.
What is it when the error has zero population mean?
its accounted for by including the constant in the equation. When the error has a zero population mean, it means that the average of all the error terms across the population is zero. This is a fundamental assumption in regression analysis and econometrics to ensure the model is unbiased.
What is uncorrelated error terms?
refers to a situation in statistical modeling or econometrics where the error terms (also called residuals) of a model are not related to each other. This concept is critical in regression analysis and other forms of statistical modeling because it ensures that the model’s assumptions are valid, leading to reliable results.
