Application of Multilabel Learning Using the Relevant Feature for Each Label in Chronic Gastritis Syndrome Diagnosis

Algorithm 1

REAL algorithm.

Step 1: The best feature subset is extracted for each label based on feature selection.

Step 2: Search for the k nearest neighbors based on the distance between two

training instances for each labels, which is calculated in the corresponding feature

subspace instead of the whole feature space.

Step 3: Calculate the posterior probability with the k nearest neighbors and further

the confidential threshold value.

Step 4: Estimate the posterior probabilities of test instances.