Predicting lncRNA–protein interactions through deep learning framework employing multiple features and random forest algorithm

Table 6 Performance of LPI-MFF and other previous RPI prediction methods on RPI1807 and NPInter

Dataset	Method	ACC (%)	SEN (%)	SPE (%)	PPV (%)	F1 (%)	MCC	AUC (%)
	RPITER	96.87	97.94	95.54	96.50	95.31	0.9369	99.29
	IPMiner	96.80	96.51	97.82	95.56	94.87	0.9350	96.61
RPI1807	EDLMFC	93.35	96.62	83.71	94.60	95.59	0.8225	96.89
	lncPro	47.34	44.51	50.62	53.24	51.23	− 0.049	50.64
	LPI-MFF	97.60	95.72	99.49	99.47	95.27	0.9755	99.57
	RPITER	95.35	98.02	92.67	93.05	94.01	0.9083	98.56
	IPMiner	95.70	95.64	94.77	95.66	95.89	0.9140	95.77
NPInter	EDLMFC	96.14	97.19	92.13	93.63	94.35	0.9135	98.59
	lncPro	50.84	73.92	27.60	50.56	48.76	0.0170	51.72
	LPI-MFF	97.67	97.58	94.83	93.35	94.41	0.9192	98.81

Bold values represent the maximum value of the corresponding evaluation indicator

ISSN: 1471-2105