Empirical Study on Indicators Selection Model Based on Nonparametric -Nearest Neighbor Identification and R Clustering Analysis

Table 2

The result of nonparametric -nearest neighbor discriminant model.

Serial number	Index	The accuracy of all the indicators	The identification accuracy after removing the th index				The degree of influence on the accuracy of the default state
			Nondefault customer’s judgment accuracy	Low default loss rate customer’s judgment accuracy	High default loss rate customer’s judgment accuracy	All customer’s judgment accuracy
1	Debt Asset Ratio	= 54.17%; = 0.00%; = 99.16%; = 97.2127%	99.16%	0.00%	54.17%	97.2127%	0%
7	Equity Ratio		99.16%	0.00%	54.17%	97.2129%	0.0002%
14	total liabilities net cash flow		99.28%	0.00%	54.17%	97.3285%	0.1158%
21	EBITDA		99.16%	0.00%	54.17%	97.2127%	0%
27	liquidity turnover rate		99.16%	0.00%	54.17%	97.2127%	0%
48	marital status		99.16%	0.00%	58.33%	97.3287%	0.116%
51	gender		99.16%	0.00%	50.00%	97.0963%	−0.1164%
64	GDP growth rate		99.04%	0.00%	54.17%	97.0968%	−0.1159%
65	CPI		99.04%	0.00%	54.17%	97.0968%	−0.1159
68	arrived pledge score		99.16%	0.00%	54.17%	97.2127%	0%