In: Statistics and Probability
You are now the manager of your favorite grocery store. Your job is to improve the store aesthetics and products to make sure that more shoppers come in to shop. You are asked to come up with a regression analysis to answer this question:
What are the differences between how you would interpret the normal regression and logistic regression when it comes down to (i) fit of the model, (ii) results of the regression (IV)
i) When it comes to fit of the model, normal regression fits a continuous linear line through the given data. Hence, the response is continuous and linearly varying with the predictor variable. On the other hand, logistic regression fits the probability of some event (success/failure, etc.) as a function of a linear combination of the predictor variables. Hence, the response here is discrete (often binary, i.e. 0/1) in nature.
ii) While linear regression gives a continuous variable as the output in response to the predictor variables, logistic regression gives the probability of occurrence of an event in binary format, i.e. 0 or 1. For example, in the given problem, we might want to predict if a customer will buy a product or not, hence we will model this probability as a function of different variables such as pst sales record. existing demand of the product, day of the week, seasonal trends, etc. Hence, the output of logistic regression is a response variable which is categorical in nature, while that of linear regression is continuous in nature.