Transcript:
This video is to show you how to do. Auto regression without stat Pro. Um, okay, so I’ve just got a random data set here and I I’m gonna do Auto regression on it. Now, one thing you’re gonna miss without stat? Pro is just showing you really briefly here. Um, you’re gonna be missing checking. How many legs were significant? So in this correlation covariance’s it shows you how many legs are significant if you don’t have set? Pro you can’t test that. So let’s just assume we need to generate three, just like in your final project. I just told you to make three legs essentially, or you’re just making three, so let’s just make those. Okay so to lag your data like one. Your data’s just bumped down one time step light to flag by two time steps, leg three is lagging by three time steps. Okay, and you just believe the bottom ones here for now. Um, okay, so you just bump your date it down. That’s what the lagging is. And now if you want to perform, what’s called? Auto regression. What you do is you just use these guys as the X variables and the original value variable here as the Y. Okay, now, one thing that data analysis might not, like are all of these blanks. So what I’m gonna do is I’m going to grab the data starting from the first non blank row, So all of these rows here have blanks in them. You can’t use those in the regression. So just start copying the data from the first complete full row. Okay, it’s that make sense, and then there we go. Oh, sorry, just go grab the titles here. Paste them across. And now you’re ready. So this is the data that I’m going to use to feed and to do my auto regression, and now what we’re gonna do, we’re going to start by using all of the legs as their Xs and then we’ll see from there. So without using stat Pro, we could also use the data analysis Toolpak. I just go grab regression. So auto regression is just regression using the legs as the X variables. So our Y is our original data. Our Xs are these three lags again without any of the blanks, so we’re just starting from q4 of 2010 here and just labels in the top row. You can include your confidence level if you want and I’m just gonna put this on a new worksheet. Call it regression. One spit up the residuals click. OK, okay now! If you look at your p-value’s, the p-value for leg three is no good, so let’s rerun a regression. Without that guy back to our data here. It’s gonna be in that tab. I’ll post up all these solutions as well back to the data tab. We’re going to rerun our regression, and now we’re not gonna use leg three, so we’re just going to go up to column J, So we’re just going to grab Legs. 1 & 2 for X variables still keep the total as or Y. Yeah, let’s put this on a worksheet. Let’s call that regression -. Oh, sorry, ok, gorgeous. Both both are p-values are very good now. These guys let’s check are just at our squirt. It’s very good as well. Okay, let’s see if it changed. Much between regression. 1 & 2 is 0.99 Oh, 3/5 and point nine nine. Oh, seven. So it got better in the second regression our significance. F also got better. So our regression. We would want to use. Is this model right here? Okay, and then you can go on from there and make your your equation as well. If you want it, it would just be that Y equals or Y of T if you will equals. Oh, sorry that five fifty nine, seventy four point two eight one plus 0.5 Oh, three to five three times. Y t minus one or like one. If you will. And then plus zero point, four, nine, seven, five, six, eight times y of T minus two or the leg to Y value. The book will step you through this as well. The piece. You just need to get to without stat. Pro is making this regression and doing these lags.