Thankyou very much for this crisp explanation. Really loved it. I wanted to ask you, in case of object detection algorithms like YOLO, how do we work with True Negatives? That number can be very big when we're working with 1000s of classes. Kindly let me know.
Hi @Murali Kumar ! Actually, go back to 3:03 to understand what I mean: F1 is a harmonic mean between Precision and Recall. But, if you develop the expression, some magic happens algebraically and the denominator will contain the arithmetic mean between FP and FN. That's the arithmetic mean I referred to at 6:44 😃
Oi Lucas. Muito obrigado! Espero que tenha visto que o vídeo inclui legendas em PT-BR. De qualquer modo, o plano é manter o áudio em Inglês para ter mais abrangência, enquanto a legenda em PT-BR presente em todos os vídeos garante que o meu público mais querido possa acompanhar e participar 🙂
Thanks man... now I got it!!! Congrats for the video!
So happy that it was useful 🙂
Great video! The graphs in the end really helped.
Crystal clear explanation. Thanks!
Thank you for such a neat explanation of F1 score!
Wow! amazing explanation. Thank you! 🙏
Thanks for the explanation!
nice presentation ..greetings from IRAQ
Thankyou very much for this crisp explanation. Really loved it. I wanted to ask you, in case of object detection algorithms like YOLO, how do we work with True Negatives? That number can be very big when we're working with 1000s of classes. Kindly let me know.
at 6:44, dont you mean harmonic mean?
Hi @Murali Kumar ! Actually, go back to 3:03 to understand what I mean: F1 is a harmonic mean between Precision and Recall. But, if you develop the expression, some magic happens algebraically and the denominator will contain the arithmetic mean between FP and FN. That's the arithmetic mean I referred to at 6:44 😃
@@DataScienceBits ok, thank you. I understand now. This was definitely a very informative and helpful video.
Why does the F1 score does not exceed 0.7 especially when the data is imbalanced and output probability is 0.5? Can you please clarify, thanks
Was here for Formula One score, life happens
TH chart is awesome. When are you publishing an r package so i can see my data in it?
Hi! All Python scripts to generate the plots can be found at github.com/felipepenha/data-science-bits
Thank you Sir!!!
adorei o video! vice devia gravar em portugues tambem!
Oi Lucas. Muito obrigado! Espero que tenha visto que o vídeo inclui legendas em PT-BR. De qualquer modo, o plano é manter o áudio em Inglês para ter mais abrangência, enquanto a legenda em PT-BR presente em todos os vídeos garante que o meu público mais querido possa acompanhar e participar 🙂