MATLAB for Machine Learning

Replacing the missing value

The next step will allow us to replace the missing value indicators. Then, we will clean the data so that the missing values indicated by the -19 code have the missing numeric value matrix provided by MATLAB, which (as mentioned) is NaN. To do this, we can use the standardizeMissing() function, which replaces the values specified in parentheses with standard missing values in an array or table:

>> SampleData = standardizeMissing(SampleData,-19);>> summary(SampleData(:,3))Variables:    age: 12×1 double        Values:            Min            15             Median         25             Max            100            NumMissing     2       

We may notice that -19 is no longer present, while the missing values are now 2 instead of 1. At this point, we can create a new table, SampleDataNew, and replace the ...