Feature Genes Selection Using Fuzzy Rough Uncertainty Metric for Tumor Diagnosis
Algorithm 1
Feature genes selection based on the rough uncertainty metric.
Input: rough decision information system , parameters α and β (α is fuzzy neighborhood radius and β as constraint to tolerate noises in data)
Output: one set of feature genes
Step 1. Initialize
Step 2. , calculate the fuzzy similarity matrix
Step 3. According to Definition 1, calculate the rough decision
Step 4. According to the definition of fuzzy neighborhood granule and Definition 4, combine with the rough decision to calculate the conditional entropy of decision attribute set D with respect to A
Step 5. , calculate , and select the attribute a that maximizes the value of . Then,
