Skip to main content
. 2022 Dec 28;13(1):95. doi: 10.3390/diagnostics13010095
Algorithm 4. Never-Ending Image Learner for image prediction.
input; an ontology-O
output; trusted instances for each group
share initial image data
for k = 1, 2…∞
           for each group ϵ O
                     extract new image data
                     filter patients
                     train data classifiers
                     assess the patient using a trained classifier
                     promote highest-confidence patient
           end for
           Share items
end for
for each class
          if X mutually exclusive with Y
                     Y ← negative instance of image
          end if
          if (Y(X) ← X)
                     Y ← trusted item
          end if
          if (co-occur ← two trusted patterns) & (co-occur ← any –Ve pattern in the same web)
                     NELL then filters out
          end if
end for