The feature splitting technique used to split the column to extract more data from the column it is the part of feature construction it doesn't have mathematical rules eg: There are scores of students and we are splitting then as 0-35 one column and 35-80 one column and 80 - 100 one column
the technique used to collab the column to improve the accuracy of the model is the part of feature engineering
eg: There are scores of students
and there are 2 columns as writing score and reading score we can collab them and create a new column as grammar and we can name them as 'C' means low marks and 'B' means mid-range marks and 'A' means high range marks