Rough Sets Data Analysis in Knowledge Discovery: A Case of Kuwaiti Diabetic Children Patients

Algorithm 2

Reduct generation algorithm.

Input: information table (ST) with discretized real valued attribute.

Output: reduct sets 𝑅_final={𝑟₁∪𝑟₂∪…∪𝑟_𝑛}

1: for each condition attributes 𝑐∈𝐶 do

2:  Compute the correlation factor between 𝑐 and the decisions attributes 𝐷

3:  if the correlation factor > 0 then

4:   Set 𝑐 as relevant attributes.

5:  end if

6: end for

7: Divide the set of relevant attribute into a different variable sets.

8: for each variable sets do

9:   Compute the dependency degree and compute the classification quality

10: Let the set with high classification accuracy and high dependency as an initial

  reduct set.

11: end for

12: for each attribute in the reduct set do

13: Calculate the degree of dependencies between the decisions attribute and that

  attribute.

14: Merge the attributes produced in previous step with the rest of conditional

  attributes

15: Calculate the discrimination factors for each combination to find the highest

  discrimination factors

16: Add the highest discrimination factors combination to the final reduct set.

17: end for

18: repeat

  statements 12

19: until all attributes in initial reduct set is processed