Рет қаралды 3,029
Want to take a class with me? Visit simplistics.net and sign up! See you there!
*Technical side note: zero-inflated simply means you have a distribution with an abnormal number of zeroes. To model zero-inflated data, it's common to use models that combine logistic regression (to predict the 0 vs 1) and logistic regression (to predict 1+). So it's technically incorrect to say that a distribution that has this characteristic (combining yes/no with degree) is "zero-inflated." In other words, zero-inflated describes the distribution, not how the distribution is modeled. Make sense?