The formula to compute the slope coefficient is: Cov(y,x) / Variance (x) ? The formula to compute the intercept is…I’m lost. Can someone please help me?
the slope coefficient is cov(x,y)/variance(x) yes intercept if you think about your normal Ybar = b0 + b1Xbar, b0 = Ybar - b1Xbar best problem to work this stuff all out- pg 270, problem # 8 of the CFAI text. 8D does the intercept.
Note: Bannisja beat me to the punch. So, we explain the same concept, but in slightly different ways. The intercept equation is (Mean of Y) - (Slope)*(Mean of x) The reason for this is that in an OLS regression, the fitted line goes through the point (mean of x, mean of y). Let’s assume that the slope is 2 and the means of X and Y are 3 and 8 respectively. So, the line goes through (3,8). If the slope is 2, for every unit decrease in X, Y decreases by 2. The intercept is the Y value when X is zero. So, going from the point (3, 8) to the point where X has a value of zero represents a decrease in X of 3. Since the slope is 2, the decrease in Y must be 2*3 = 6. So, the intercept is 8 - 2*3 = 2. , and the regression equation will be Y = 2 + 2X
busprof Wrote: ------------------------------------------------------- > Note: Bannisja beat me to the punch. So, we > explain the same concept, but in slightly > different ways. > > The intercept equation is (Mean of Y) - > (Slope)*(Mean of x) > > The reason for this is that in an OLS regression, > the fitted line goes through the point (mean of x, > mean of y). > > Let’s assume that the slope is 2 and the means of > X and Y are 3 and 8 respectively. > > So, the line goes through (3,8). If the slope is > 2, for every unit decrease in X, Y decreases by 2. > The intercept is the Y value when X is zero. So, > going from the point (3, 8) to the point where X > has a value of zero represents a decrease in X of > 3. Since the slope is 2, the decrease in Y must > be 2*3 = 6. > > So, the intercept is 8 - 2*3 = 2. , and the > regression equation will be Y = 2 + 2X Bannisja: hope the studies go well, girl. You’ll get it this time, I’m sure.
i plan to get it- have barely started but will turn it on as it gets colder and colder in our fine city. you’ve motivated me- will get back to my SS12 problems now, on #8 and said i’d finish these today. football is soon though, time to get crackin’. good explanation above.
Thanks y’all. Right now I’m studying with 2008 textebooks which I borrowed from local CFA society, will probably get 2009 in next week or two. Yes, bannisja, I am on question 8D from Reading 11, got to the regression question and got stumped. I wasn’t prepared to calculate the slope and intercept, but now I’ll add it to my list of formulas to memorize. Thanks again for all who responded
rhythm- quant is 100% the exact same 2008 and 2009- i have both text being a failer in 2008. so the good news, the texts you’re using now won’t change at all at least for quant and you aren’t missing anything in the 2008 version. i actually circled that problem as one of the harder ones in SS11, so you weren’t alone getting stumped the 1st time around.
bannisja, thanks for the feedback. I wasn’t ready to go back and forth from covariance to variance, either. L2 seems to unfold like an onion. At first it doesn’t seem so tough, then the layers of the onion just keep unfolding.