There are basically two types of problems that predictive modeling deals with:
Classification problems
Regression problems
Classification
In some cases, we want to predict which group an observation is part of. Here, we are dealing with a quality of the observation. This is a classification problem. Examples include:
The prediction of the species of plants based on morphological measurements
The prediction of whether individuals will develop a disease or not, based on their health habits
The prediction of whether an e-mail is spam or not
Regression
In other cases, we want to predict an observation's level on an attribute. Here, we are dealing with a quantity, and this is a regression problem. Examples include:
The prediction of how much individuals will cost to health care based on their health habits
The prediction of the weight of animals based on their diets
The prediction of the number of defective devices based on manufacturing specifications