In this context, an encoding technique is a method of converting the values of a dataset or of an attribute into a dataset so that data analysis techniques and machine learning models can process them more easily. One-hot encoding is a method that is to be used on categorical data.
Let's discuss the theory of one-hot encoding first. Say we have a simple table with the following data:
Sample tabular – categorical data
Now, the string data in the City column is not very machine-friendly. There are, of course, a number of machine learning models that will have no problem processing this attribute (for example, random forests), ...
