MGEC11H3 Lecture Notes - Lecture 8: Normal Distribution
Document Summary
Inferences inmultiple regression y potb prints assume. If the probability or close to a straightling plot is a straight line this assumption is met. Themodel is homoskadast. ie var 2 model is hetroskadastic result error. From multiple regression model usethe sample abtain the residuals. 2 dependent variable and original h deep variables as the explanatory variables. Regression result shows fe 466 p value ooo2. The modelviolates one assumption that e has a constant variance. 6 taking transformation when a model violates some assumptions sometin a transformation problem a commonly used log transformation on the variables may the transformation is the sue. Regress log price on og ut size sqft bedroom colonia. F statistic has a smaup rdue. modelissignifiaprob. plot. is almost a straight line so enurmu. Hi var elt constant constant hung variables do not log. Conclusion model is homoskadat. ie hence the model price on bedrooms lotsize lsqfhi should be used insteadof the original variables.