Logistic Regression (page 2 of 6)

The logistic regression algorithm uses the logistic function to calculate the probabilities, which are converted to odds, odds ratios, and the log of the odds ratios (logit function). The challenge with probabilities is that they are bounded between 0 and 1 and they are non-linear (i.e., going from 0.1 to 0.2 doubles the probability but going from 0.8 to 0.9 barely changes the probability). As a result, we cannot use a standard linear regression algorithm when our target is categorical.