Hey HN!<p>I've been deep-diving into an intriguing question while working on my side project and during my studies.<p>I experimented with predicting the length of my future phone usage sessions (from unlock to lock), and the outcomes have been pretty remarkable.<p>I got my usage data in format<p>timestamp_of_session, duration_in_seconds<p>By applying a straightforward linear regression model, I achieved a Mean Squared Error of just around 3 minutes. Isn't that fascinating?<p>I'm really excited to hear your thoughts and insights on this! Thanks in advance for your input!
How do you think ad targeting works? First google and friends build a model of your digital habits and then they figure out the best ways to intervene in the digital space with ads that you will click on. Same with Facebook. These companies make billions every day and that's because they are very effective at using AI (linear regression) to modify user behavior on a mass scale and increase revenue for the advertisers.