This document describes the MIL (Multiple Intervals List) algorithm for discretization of continuous attributes for classification problems. It involves 4 scans of the training data: 1) calculate min and max, 2) calculate thresholds, 3) calculate optimal merged sub-intervals, 4) discretize attributes. An example shows applying it to a CGPA attribute to classify students into grade categories. The algorithm merges sub-intervals until their combined instances exceed 1/3 of the threshold. It has linear time complexity and requires only 4 data scans, making it efficient compared to other discretization methods.