How to optimize a machine learning model for detecting phishing tweets with minimal false positives?
I’m currently working on a machine learning project aimed at detecting phishing tweets. While my model performs reasonably well, I’m facing challenges with a relatively high number of false positives, which reduces the overall accuracy and usability of the system. Here are the details of my current setup: