How to assign numbers to categorical values in a dataset?

I'm preparing a dataset for machine learning. The dataset contains a column name "Holiday". The column contains more than a million row of values. It is categorical in nature and contains 4 unique values - 0 (as a string), a, b, c.

I want to assign the values 0 to 0 and 1 to the rest of them - a, b and c. How do I do that? Is there a readymade function?

Answer 1

[Adam Danz]

If you want to return logical values,

dummyVars = Holiday ~= '0';  % Holiday is categorical

If you want to return integer values,

dummyVars = double(Holiday ~= '0'); % Holiday is categorical

Note that any value of Holiday that doesn't equal 0 will be assigned a value of 1.

[pp]

How would i apply this over the entire variable (over 1 million rows)?

[Adam Danz]

I've updated my answer. The previous version treated "Holiday" as strings. The current version treates "Holiday" as categorical. The output is a vector the same size as the input.

[pp]

Thanks! That did the job. Is it possible to extend this so that we can assign other numbers to a, b and c? Let's say 1, 2 and 3?

[Adam Danz]

In that case, you can use

[groups, groupID] = findgroups(Holiday)

or

[groupID, groups] = grp2idx(a); % requires stats & ML toolbox