When dealing with the choice of the right set of models to experiment with to solve your problem, you should bear in mind the following classification of modeling problems that will help you take your search in the right direction:

• Clustering: These are the problems where you need to regroup your data based on common features, that is, measures of similarity. For instance, you may need to cluster your customers to perform further analyses on their shopping behavior.
• Classification: Within this kind of problem, you need to define a rule able to assign new elements to one of the categories available within your population, given a set of features. You may, for instance, be wondering about defining the category of products a new customer will most probably like, based on some personal information you have about them.
• Regression: You are facing a regression problem when you need to understand how different ingredients contribute to a final outcome. A typical example could be a product you need to advertise, for which you would like to know on which advertising platform you should concentrate the main part of your budget. Imagine having a dataset showing, for a group of products, how much budget was placed for each advertising platform, and the final revenue obtained from that product. Modeling the relationship between the level of investment on each platform and the final revenues would be considered a classic case of regression.