International Journal of Educational Technology in Higher Education

Table 7 Comparative analysis of classification models applied to data enhanced with graph concepts

From: Extracting topological features to identify at-risk students using machine learning and graph convolutional network models

Metrics	Data used	ACC	\(\Delta _{ACC}\), %	AUC	\(\Delta _{AUC}\), %
Original dataset	DS	0.845	Ref.	0.948	Ref.
Original dataset	DS+GE	0.86	+1.775	0.976	+2.954
Cosine metric	DS+GF	0.873	+3.314	0.967	+2.004
	DS+GF+GE	0.874	+3.432	0.970	+2.321
	GCN (DS+GF)	0.882	+4.379	0.972	+2.532
Euclidean metric	DS + GF	0.847	+0.237	0.957	+0.949
	DS + GF + GE	0.863	+2.130	0.975	+2.848
	GCN (DS + GF)	0.850	+0.592	0.957	+0.949

\(\Delta\) shows the boost in performance compared with original dataset (Ref.)
DS: initial dataset; GF: graph features; GE: graph embeddings

Back to article page