Hi, many thanks for your great video series. Just a quick question: Shouldn't the the numerator (x bar) be 30^2 in the equation at min 2:40 be sqrt(1/5 + 30^2 / 1000.0) ~ 1.05? Because if we don't square it, then the result differs. Many thanks in advance. Great videos.
Hi, shouldnt SSE = sum of squared diff between Actual values - Estimated values(y^) instead of expected values. expected values means average of y hat. Correct me if I am wrong. thanks
Hi, it's the matter of using terms. actual value is yi, and expected (or estimated) value is ŷ. Therefore, yi - ŷ is the error. ŷ indicates the regression line, while yi indicates data points. Therefore, the difference between actual data points and regression line is the error. If all data points are exactly on the regression line, it's y = x, and there are no errors. To calculate SSE and SSR, the average of ŷ is not usually considered. This is sum of squared for simple linear regression. SST = SSR + SSE; Σ (yi - ȳ)2 = Σ (ŷ - ȳ)2 + Σ (yi - ŷ)2. I recommend this article. agronomy4future.org/archives/18092. Please let me know if you have further questions.
the standard error when using one-way ANOVA is calculated as the pooled standard deviation / √n. For the details, please see this post, agronomy4future.org/archives/15708
At 2:41, It should be sqrt(1/5 + 30^2 / 1000.0), not sqrt(1/5 + 30 / 1000.0). It's a typing error.
A great explanation. Thank you!
Hi, many thanks for your great video series. Just a quick question:
Shouldn't the the numerator (x bar) be 30^2 in the equation at min 2:40 be sqrt(1/5 + 30^2 / 1000.0) ~ 1.05?
Because if we don't square it, then the result differs.
Many thanks in advance. Great videos.
Thank you for your feedback. You're absolutely correct. It's a typing error. It should be sqrt(1/5 + 30^2 / 1000.0). Thanks!!
Hi, shouldnt SSE = sum of squared diff between Actual values - Estimated values(y^) instead of expected values. expected values means average of y hat. Correct me if I am wrong. thanks
Hi, it's the matter of using terms. actual value is yi, and expected (or estimated) value is ŷ. Therefore, yi - ŷ is the error. ŷ indicates the regression line, while yi indicates data points. Therefore, the difference between actual data points and regression line is the error. If all data points are exactly on the regression line, it's y = x, and there are no errors. To calculate SSE and SSR, the average of ŷ is not usually considered. This is sum of squared for simple linear regression. SST = SSR + SSE; Σ (yi - ȳ)2 = Σ (ŷ - ȳ)2 + Σ (yi - ŷ)2. I recommend this article. agronomy4future.org/archives/18092. Please let me know if you have further questions.
It is great. Is there any calculation difference if apply to one way ANOVA?
the standard error when using one-way ANOVA is calculated as the pooled standard deviation / √n. For the details, please see this post, agronomy4future.org/archives/15708
Anyong! When's the next one coming out?