An Empirical Study on Downstream Dependency Package Groups in Software Packaging Ecosystems

Table 10

Prediction results of CDDG size.

Prediction models		CDDG size  100			CDDG size  100
		Cargo	CPAN	RubyGems	Cargo	CPAN	RubyGems
Linear regression	MAE	35.14	35.53	101.74	12.16	14.43	13.39
	RMSE	45.77	61.84	117.75	14.71	22.44	24.10
	MAPE	13.08%	18.78%	72.77%	32.76%	37.66%	25.99%
Random forest	MAE	25.11	37.07	61.23	4.53	1.41	11.79
	RMSE	33.05	65.41	87.47	6.35	5.41	15.80
	MAPE	9.11%	17.18%	36.86%	10.65%	2.4%	25.18%
KNN	MAE	75.69	44.81	67.61	13.97	12.65	13.08
	RMSE	95.61	74.11	93.80	18.55	15.77	16.90
	MAPE	32.09%	20.10%	37.98%	41.29%	30.26%	30.96%
AdaBoost	MAE	75.48	40.14	99.42	5.59	3.32	14.56
	RMSE	116.62	68.91	150.93	7.02	5.88	18.03
	MAPE	19.53%	19.80%	39.29%	13.47%	7.2%	30.65%
GBRT	MAE	23.27	36.23	90.33	4.51	1.20	12.30
	RMSE	28.40	61.69	157.87	6.36	3.58	16.86
	MAPE	9.20%	17.62%	35.66%	10.59%	2.2%	26.48%

Highlight the optimal results under each indicator of different models.