Replies: 2 comments 4 replies
-
|
I don't think the NBA game changes much in the course of a few months, or even a few years... teams might get better or worse, but this is reflected by updated stats. The essence of the game might slowly change, not quickly.... 3 point shots have become more valuable over the last 5 years... but if you look at a game 2 years ago, compared to a game yesterday, it seems very similar to me. Also one of the main things to look at when evaluating a model is how well it generalizes to test data never seen before, and not overfitting. A model that cant generalize to future games, is kinda missing the point.... I have also specifically tested this theory, by stopping training up to a full season early, then using the last 250 games or so as testing, leaving over 1000 game gap between training and testing, It isn't detrimental and docent destroy accuracy on its own. Its more harmful to just not have those games or 1000 less games to train on, but is negligible if you added another 1000 games to the beginning of the dataset. |
Beta Was this translation helpful? Give feedback.
-
|
I have the same question
Em sexta-feira, 24 de março de 2023 às 08:46:35 BRT, jidemaniax ***@***.***> escreveu:
Please i have been wondering how to train the model with new data further down the line
—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>
|
Beta Was this translation helpful? Give feedback.
Uh oh!
There was an error while loading. Please reload this page.
-
I have seen the Dataset files, and I have found a file called "DataSet-2022-23.xlsx" when I opened it, it had all the data of the matches, but I have reviewed the file well, and they only have the matches up to date 2022 -06-16, so that worries me since if you only have data up to that date, the predictions could be less accurate and I even dare to say that the predictions would no longer be valid, would they?
Beta Was this translation helpful? Give feedback.
All reactions