[Retracted] A Deep Learning Approach for a Source Code Detection Model Using Self-Attention

Table 3

Classification performance depending on AST splitting methods (percentage).

Description	F1 score (BigCloneBench)	F1 score (OJClone)
AST-full (model 1)	75.8	68.9
At-BiLSTM (model 2)	94.3	95.8
AST-block (model 3)	93.2	91.5