Prediction of Flight Time Deviation for Lithuanian Airports Using Supervised Machine Learning Model

Table 4

Departure flights’ estimation results (expressed as a percentage).

Algorithms	Dataset	Measures (%)
		Sensitivity/recall	Precision	Specificity	F measure	Accuracy
Probabilistic neural network	Testing	96.46	86.27	26.84	90.14	86.27
	New data	94.00	80.77	10.38	86.66	82.01
Multilayer perceptron	Testing	90.90	85.65	48.15	87.86	85.65
	New data	88.04	80.76	33.52	84.07	81.80
Decision trees	Testing	91.78	89.87	71.47	90.65	89.87
	New data	89.35	87.68	69.63	88.37	87.54
Random forest	Testing	94.94	91.72	65.66	92.82	91.72
	New data	89.97	84.68	46.51	86.60	85.48
Tree ensemble	Testing	94.68	93.25	76.96	93.74	93.25
	New data	90.02	88.48	70.72	89.11	88.53
Gradient boosted trees	Testing	96.62	96.02	87.33	96.22	96.02
	New data	93.35	92.06	79.18	92.52	92.63
Support vector machine	Testing	99.86	82.88	0.70	90.53	82.88
	New data	99.82	83.04	2.17	90.63	83.07