>Previous data-driven studies on incel-related communities have aimed to identify key words, to make inferences about incel demographics, to characterize activ

>Previous data-driven studies on incel-related communities have aimed to identify key words, to make inferences about incel demographics, to characterize activity on various platforms, and to automate detection of related communities Do naive-bayes based spam filters not work any more? Or just no one knows how to implement them?