Data Types and Level of Measurement in Machine Learning
Types of Data
There are several debatable categorization methods of data.
The first widely spread theory, or level of measurement, is by S. Stevens. The theory categorizes data into four types, nominal, ordinal, interval, and ratio.
Other methods are proposed for other fields of research. For example, N. R. Chrisman proposed a different method for cartography. However, these are not generic enough for data science. They are more general than a specific field of research.
For machine learning, many statistical data types have been proposed. Some examples of data types and their relations with the level of measurement are shown in the following chart.
L Ma (2020). 'Data Types and Level of Measurement in Machine Learning', Datumorphism, 01 April. Available at: https://datumorphism.leima.is/wiki/machine-learning/feature-engineering/data-types/.