ECOR 1010 Lecture 14: Lectures 14 15 Summary Correlation and Regression Page 2
Document Summary
Good correlation means regression can be performed to predict y"s from x"s. If this assumption is not met the regression will still produce values for the slope and intercept, but these will not be interpretable. Regression is a process where a mathematical model is fitted to a set of data tests whether model provides reasonable description of observations: give confidence that predictions we make with the model will be good. Regression line: straight line that best predicts y (symbol: y hat: b is y intercept (y hat when x = 0, slope is difference in y hat with a change of x by 1 unit. Residuals or deviations or errors are deviations between observed and calculated y values from equation.