Identification of Enterprise Financial Risk Based on Clustering Algorithm

Table 3

The screening effect of K-means clustering on financial risk.

years	High-risk cluster proportion (%)	Number of high-risk companies (homes)	Y + 0 year recall rate (%)	Y + 1 year accuracy rate (%)	Y + 2 year accuracy rate (%)	Y + 3 year accuracy rate (%)	Y + 4 year precision rate (%)
2008	8	1 18	36.0	10.2	17.0	20.3	20.3
2009	2.4	38	17.2	23. I	23	23	23、
2010	6.2	105	20.5	10.5	14.3	15.2	16.2
2011	6.0	123	31.3	7.2	8.9	11.4	16.3
2012	5.5	126	19.2	6.4	7.9	9.5	12.0
2013	6.5	157	22.7	8.3	10.8	13.4	16.6
2014	21.0	515	55.6	6.6	9.7	12.4	16
2015	16.4	784	20.5	4.3	7.1	10.0	12.6
Mean value	9.0	246	27.9	9.6	12.4	14.4	16.6