Awesome video! Quick question: since the derivative of the harmonic mean has that squared term, does that mean that the harmonic mean can only ever increase as it 'sees' additional samples? This doesn't make intuitive sense to me since the derivative should be able to be negative to allow for a given data point to 'bring down' the mean (e.g. a terrible recall score will drag down an F1 score even when precision is perfect). I'm out of my depth in trying to check your work on the derivative but let me know if I'm missing something!!
This is fantastic. You are so skilled at explaining why mathematical concepts are important in data science. I learned so much from this video!
Thanks!
Awesome video! Quick question: since the derivative of the harmonic mean has that squared term, does that mean that the harmonic mean can only ever increase as it 'sees' additional samples? This doesn't make intuitive sense to me since the derivative should be able to be negative to allow for a given data point to 'bring down' the mean (e.g. a terrible recall score will drag down an F1 score even when precision is perfect). I'm out of my depth in trying to check your work on the derivative but let me know if I'm missing something!!
clean, quick, fun. You're a talented explainer
None of these concepts beyond the equations were fleshed out in any of my data science courses. Thanks a bunch!
that "let s take a look" at 6:07 sounded strangely like Luis SERRANO :)
Thanks 👍🏽
Was just learning this a couple months ago studying for advanced calculus, interesting topic, nice vid.
This is a great explanation! Thank you.
awesome, thank you
You bet!
Amazing we need more
thank you
Welcome!