Problem of Discrete Values

During the real use of perSimplex product cases happen when the input data contain values, which are not similar mutually. Various questionnaires and polls are the typical example, which the questions with non-similar possible answers include in.

The marking of columns with discrete values is executed within the perSimplex product directly in the first row of the CSV input file. The columns with discrete values contain the character ‘#’ in angle brackets, i.e. the [#] string in the column name, without spaces and other characters. The aftermath of the marking of the column with discrete values is that the SimplexDivide application estimates the different values in the given column as absolutely non-similar (zero degree of similarity) while evaluating the values similarity. Said in other words, the values similarity degree in these columns is either 0-different or 1-identical.