Statistics for Data Science

Common guidelines

Guidelines for classification will most often adhere to the following:

• There should not be any doubt about a definition of the classes

• All the classes should preferably have equal width or length

• The class-limits should be selected in such a way that no value of the item in the raw data coincides with the value of the limit

• The number of classes should preferably be between 10 and 20, that is, neither too large nor too small

• The classes should be exhaustive, that is, each value of the raw data should be included in them

• The classes should be mutually exclusive and non-overlapping, that is, each item of the raw data should fit only into one class

• The classification must be suitable for the object of inquiry ...