Table 5 Results of ablation studies on different programming language datasets.

From: Difficulty aware programming knowledge tracing via large language models

Methods

C++_AUC

C++_ACC

C_AUC

C_ACC

Java_AUC

Java_ACC

Python_AUC

Python_ACC

DPKT w/o LLM

74.8

73.2

74.3

70.1

74.0

71.4

75.5

66.9

DPKT w/o BERT

73.5

72.1

73.1

69.0

73.0

70.5

74.1

65.4

DPKT w/o GAT

75.2

73.8

74.5

70.8

74.7

71.9

75.9

67.2