Thanks a lot, Helped me understand mutiple regression very well, What is the basic difference between a univariable and linear regression analysis?? Why didnt I find u earlier😭
Hey Brandon.. Thanks for such an awesome series on Statistics. Could you please also make a video on the venn diagram that you did in the previous video? All the estimators in Linear Regression could be very confusing. :) If I were to summarise it for self understanding: Mean Squared Error is practically the mean of the unexplained error(residuals) rmse = standard error mse(unexplained error) = sqrt(mse) The standard deviation of the predicted values is dependent on the deviations in the independent variable. which is why the slope itself could have its own distribution and thus its own standard error calculated with the help of rmse, with the formula: rmse * sqrt(1/n + ( (x-mean(x)^2) / sum(x-mean(x)^2) ) ) - the nasty looking formula. :P and thus they have their own confidence intervals: which uses the predicted values (yhat) instead of x, since we are calculating all these estimators based on our predictions. s_squared_predicted = mse + (sd_pred: nasty formula)**2 The standard error for these variations is: sqrt(mean_squared_error + the variation in the predicted amount) and thus the confidence intervals for the s_squared_predicted: using s_squared_predicted and yhat I have tried to simplify the various standard deviations/standard errors and the multiple confidence intervals that you described when doing the linear regression with just one variable. I hope I have it right...
Dear Brandon. I have just open your UA-cam Channel and I am amazed and at the same time worried while I only just got to know about this channel. The materials in there are priceless. I wish to view or learn about Measures of Associations like Odds Ratios and Relative Risks and I do not know which Video to open. Can you help point me in the right direction? Thanks. JC
Really great videos - thank you - can you point me one that explains what inferential statistics to use for a categorical data from one sample - just looking at relationships between variables thank you
@@BrandonFoltz Hi - thank you for replying - I am exploring relationships between 6 categorical variables - a mix of multiple responses - 1 x single item ordinal 5 point numeric rating scale, 3 x Multiple categorical ( 2 x 5 point Likert Agreement scales, both 8 items and 1 nominal dichotomous Y/N x 7 items. 1 = 5 point numeric scale x 6 items and 1 x 5 point anchored rating scale. - You are very kind - thank you.
Dear Brandon, thanks for your amazing videos ❤️ currently I am struggeling to use a moderating analysis in the multiple regression model (using R) do you happen to have any recommendations regarding that? Thanks in advance and thanks a lot for the amazing Content
Thank you for making these video's! They have become my go to for understanding what my instructors are talking about in statistician lingo.
Namaste, great video sir. ❤️
Thanks for the best teaching
beautifully explained it thank you so much
Thanks a lot, Helped me understand mutiple regression very well, What is the basic difference between a univariable and linear regression analysis?? Why didnt I find u earlier😭
Hey Brandon.. Thanks for such an awesome series on Statistics.
Could you please also make a video on the venn diagram that you did in the previous video?
All the estimators in Linear Regression could be very confusing. :)
If I were to summarise it for self understanding:
Mean Squared Error is practically the mean of the unexplained error(residuals)
rmse = standard error mse(unexplained error) = sqrt(mse)
The standard deviation of the predicted values is dependent on the deviations in the independent variable.
which is why the slope itself could have its own distribution and thus its own standard error
calculated with the help of rmse, with the formula:
rmse * sqrt(1/n + ( (x-mean(x)^2) / sum(x-mean(x)^2) ) ) - the nasty looking formula. :P
and thus they have their own confidence intervals:
which uses the predicted values (yhat) instead of x, since we are calculating all these estimators based on our predictions.
s_squared_predicted = mse + (sd_pred: nasty formula)**2
The standard error for these variations is:
sqrt(mean_squared_error + the variation in the predicted amount)
and thus the confidence intervals for the s_squared_predicted:
using s_squared_predicted and yhat
I have tried to simplify the various standard deviations/standard errors and the multiple confidence intervals that you described when doing the linear regression with just one variable.
I hope I have it right...
lol
Dear Brandon. I have just open your UA-cam Channel and I am amazed and at the same time worried while I only just got to know about this channel. The materials in there are priceless. I wish to view or learn about Measures of Associations like Odds Ratios and Relative Risks and I do not know which Video to open. Can you help point me in the right direction? Thanks. JC
Really great videos - thank you - can you point me one that explains what inferential statistics to use for a categorical data from one sample - just looking at relationships between variables thank you
Hello! Thanks for watching. Not sure I understand the question. Can you give me an example of the type of analysis you are looking to conduct?
@@BrandonFoltz Hi - thank you for replying - I am exploring relationships between 6 categorical variables - a mix of multiple responses - 1 x single item ordinal 5 point numeric rating scale, 3 x Multiple categorical ( 2 x 5 point Likert Agreement scales, both 8 items and 1 nominal dichotomous Y/N x 7 items. 1 = 5 point numeric scale x 6 items and 1 x 5 point anchored rating scale. - You are very kind - thank you.
I request you to explain how to calculate hardware statistics or computer
Can we have both categorical, ordinal variable together and run a regression model??
Yep! Very flexible. Under the hood, the software turns them all into dummy variables (or one-hot encoding).
3.21 there is something strange: it should be the "potentially removed variable" and not the "added variable"
Dear Brandon, thanks for your amazing videos ❤️ currently I am struggeling to use a moderating analysis in the multiple regression model (using R) do you happen to have any recommendations regarding that? Thanks in advance and thanks a lot for the amazing Content
Hello
Hello!