Machine Learning Algorithms for understanding the determinants of under-five Mortality

Table 2 The performance of the prediction models with all factors based on various indices for two ratios

Train/test ratios	Measures	Decision tree	Random forest	Naive Bayes	K-Nearest neighbour	Logistic regression	SVM regression	Neural network	Ridge regression
70/30	Sensitivity	93.30	72.66	40.00	58.62	64.11	63.27	80.62	71.51
	Specificity	94.74	96.49	94.32	94.10	95.67	95.70	96.47	94.39
	Precision	20.47	48.53	14.95	10.42	35.91	36.52	47.92	15.07
	Accuracy	94.72	95.46	92.99	93.69	94.52	94.48	95.86	94.08
	F1 Score	33.57	58.19	21.77	17.69	46.03	46.31	60.11	24.90
	Negative Predictive value	99.90	98.73	98.44	99.49	98.60	98.52	99.20	99.58
	Cohen’s Kappa values	0.32	0.56	0.19	0.17	0.44	0.44	0.58	0.24
80/20	Sensitivity	92.91	75.41	41.55	60.78	65.00	65.55	79.27	71.31
	Specificity	94.80	96.62	94.35	94.13	95.79	95.87	96.71	94.42
	Precision	21.61	50.55	15.75	11.36	38.10	39.38	51.83	15.93
	Accuracy	94.77	95.69	93.04	93.73	94.61	94.68	95.96	94.08
	F1 Score	35.07	60.53	22.84	19.14	48.04	49.20	62.68	26.05
	Negative Predictive Value	99.88	98.85	98.45	99.49	98.57	98.55	99.05	99.55
	Cohen’s Kappa values	0.33	0.58	0.20	0.17	0.45	0.47	0.60	0.25

ISSN: 1756-0381