thanks for the overall great vid, I learned a lot! But I have open questions: to me it is not quite clear what exactly is hiding behind the estimated breakdown point. you said to get there... one starts replacing values with extreme ones. buw how extreme? (I suppose it makes a difference if we replace a value with sth 2 or 5 SD's above mean.) also the criterium for when we count the estimation to be "broken" and when "still okay" was not talked about.
Thanks Chris.. this had to be one of the best series on the internet on intro to robust stats
Are there robust estimators for skewness and kurtosis?
You make a good teacher, you look enthusiastic about the material to the point where I think I want to be interested too
Thank you Chris for this well-presented and explained lecture.
Really excellent lecture, wish I could take all your statistic courses.
thanks for the overall great vid, I learned a lot! But I have open questions:
to me it is not quite clear what exactly is hiding behind the estimated breakdown point. you said to get there... one starts replacing values with extreme ones. buw how extreme? (I suppose it makes a difference if we replace a value with sth 2 or 5 SD's above mean.) also the criterium for when we count the estimation to be "broken" and when "still okay" was not talked about.
Thank you Chris for this very informative lecture.
Chris you're a lad