ECON 174 Lecture Notes - Lecture 5: Simple Linear Regression, Linear Regression, Regression Analysis
22 Jan 2020
School
Department
Course
Professor
Document Summary
Chapter 5: linear regression models, least squares estimation, selecting and evaluating regression models, forecasting with regression, non-linear regressions, spurious regressions, stationarity, casuality. Simplest regression model relates a single dependent variable to a single regressor: B0 is the intercept, the prediction of y1 if x1 = 0. B0 + b1xi represents part of yi that"s explained in model. I is the error, including everything that affects yi that is not captured in xi. (approx) linear relationship between yi and xi is key assumption in model. Zero mean in the errors (ols will do this automatically) No correlations between errors and independent variables. Translates immediately to a time series context: yt= b0+b1xt+ where b1 captures how an increase in time series x at time t affects our predictors of times series y at time t. Extends easily to k predictor time series: Interpretation of b0 and et stays the same.
