Open ghost opened 3 years ago
If precision is high but recall is low, it usually means the clustered results consist of many small clusters. In an extreme case, each instance is taken as a clustere. A possible reason may lie in the wrong linkage when building the graph. Checking the built graph or the training loss may help.
Hi ! I am training LGCN on some custiom dataset. Training seems to run somehow but I get the following warning.
"/home/redion/redion_files/learn-to-cluster/faceCluster/lib/python3.6/site-packages/sklearn/metrics/_classification.py:1272: UndefinedMetricWarning: Recall is ill-defined and being set to 0.0 due to no true samples. Use
zero_division
parameter to control this behavior. _warn_prf(average, modifier, msg_start, len(result)) " What does this mean ? I get 97 percent accurracy but precision and recall are 0.