Hi John, thanks for your message. Could you clarify what you mean by that equation? As it stands I am not entirely sure as to its meaning. I can then either point you in the direction of a relevant video, or will make one. Thanks, Ben
Ok. I think what 'sse' means in this context is the sum of squared errors. This is the variation in y which your model fails to explain. 'ssy' is the total variation in y. So let's say you are trying to understand which factors affect house prices (your y) in terms of an econometric model - the ssy is the total variance of house prices away from the mean. sse is the variance which your model doesn't explain. The ratio (ssy-sse)/ssy hence gives the percentage of variance explained by your model
Hello, first of all your videos are awesome and way better than all the crap taught at the university. Second, I have a question: R² amounting to zero is the line of the sample mean. However, imagine a parallel line to the mean line beneath all the data points. Shouldn't R² be zero here considering the model basically doesn't explain anything? But if I apply the definition, it isn't at all. Thanks in advance!
DoD_J4Jc3 Hmmm that's an interesting thought. The R^2 wouldn't be positive in your example, seeing as Sum of squares error for that line would be more than TSS I think that explained sum of squares must be negative. Implying a worse fit than simply using the mean as an estimator??
Glad it helped! Let me know if you have any further questions.
Let me know if this makes sense. It's quite hard to explain in writing!
watching your vids in 2022
How are you doing these days sir, I hope you're living a healthy life :)
Hi John, thanks for your message. Could you clarify what you mean by that equation? As it stands I am not entirely sure as to its meaning. I can then either point you in the direction of a relevant video, or will make one. Thanks, Ben
Ok. I think what 'sse' means in this context is the sum of squared errors. This is the variation in y which your model fails to explain. 'ssy' is the total variation in y. So let's say you are trying to understand which factors affect house prices (your y) in terms of an econometric model - the ssy is the total variance of house prices away from the mean. sse is the variance which your model doesn't explain. The ratio (ssy-sse)/ssy hence gives the percentage of variance explained by your model
Awesome video. You made me understood better than my professor taught me. Thanks.
You are truly a god amongst men
Fitted values are points on the line: they are infinite. What does it mean to sum them from 1 to N?
Hello, first of all your videos are awesome and way better than all the crap taught at the university. Second, I have a question: R² amounting to zero is the line of the sample mean. However, imagine a parallel line to the mean line beneath all the data points. Shouldn't R² be zero here considering the model basically doesn't explain anything? But if I apply the definition, it isn't at all. Thanks in advance!
DoD_J4Jc3 Hmmm that's an interesting thought. The R^2 wouldn't be positive in your example, seeing as Sum of squares error for that line would be more than TSS I think that explained sum of squares must be negative. Implying a worse fit than simply using the mean as an estimator??
can you add captions to this video?
very clear!!
best