By Cristiano
Decision tree trained with 200k documents (half true, half synth) with a max split = 3 yields a model with 91% accuracy.
Removing the most important features (n_tokens and n_sentences) yields a model with 91% accuracy nonetheless.
Jul 23, 20261 min read
By Cristiano
Decision tree trained with 200k documents (half true, half synth) with a max split = 3 yields a model with 91% accuracy.
Removing the most important features (n_tokens and n_sentences) yields a model with 91% accuracy nonetheless.