Escuchar "Top 10 Machine Learning Gotchas - Paige Bailey"
Síntesis del Episodio
Si quieres ver el vídeo con las slides: https://www.youtube.com/watch?v=q35jlAD_jQ4
"There are two styles of general overfitting: over-representing performance on particular datasets; and (implicitly) over-representing performance of a method on future datasets."
-- John Langford (Microsoft)
=======================
Is your model performing excellently during training, but failing to generate accurate predictions on real data?
=======================
Chances are, you may have fallen into one of the classic traps of overfitting! Never fear: we have you covered.
In this talk, we will walk through common examples of overfitting machine learning models, and give you a set of data analysis tools to recognize and combat the problem. No experience with machine learning is necessary; just a desire to learn, experience using a programming language (preferably R, Python, Scala, or Java), and a passion for feature engineering!
"There are two styles of general overfitting: over-representing performance on particular datasets; and (implicitly) over-representing performance of a method on future datasets."
-- John Langford (Microsoft)
=======================
Is your model performing excellently during training, but failing to generate accurate predictions on real data?
=======================
Chances are, you may have fallen into one of the classic traps of overfitting! Never fear: we have you covered.
In this talk, we will walk through common examples of overfitting machine learning models, and give you a set of data analysis tools to recognize and combat the problem. No experience with machine learning is necessary; just a desire to learn, experience using a programming language (preferably R, Python, Scala, or Java), and a passion for feature engineering!
Más episodios del podcast CODEMOTION 2017
Redux a través de la música - Elena Torró
22/02/2018
Tuenti: De una aplicación a N aplicaciones con la misma base de código - Eduardo González
02/02/2018
Android TV de 0 a 100 - Mario Velasco
22/01/2018
ZARZA Somos ZARZA, la firma de prestigio que esta detras de los grandes proyectos en tecnología de la información.