 What is the equation of bias in machine learning? Let's say that we have a linear regression model where the parameter is theta hat. For this equation, let's say the true value is theta star. And if we were to train this model, let's say 100 times, we'll get 100 estimates of theta hat. And if we average it, we might get some value over here. This distance is called the bias. And what we did was just take the average value of different theta estimates. We put that in the first term. And this theta star here is the true estimate, which is the second term. And so that's how we have the bias. If this is equal to zero, then this linear regression equation here is an unbiased estimator.