An Artificial Intelligence Approach to Financial Fraud Detection under IoT Environment: A Survey and Implementation

Table 3

F-measure of clustering algorithms before and after the feature selection process.

Ratio	Stage	EM	simpleK	FarthestFirst	XMeans	DensityBased
50:50	Before; After	0.5297; 0.6134	0.4815; 0.6443	0.2692; 0.6647	0.4815; 0.6443	0.4806; 0.6492
60:40	Before; After	0.5787; 0.6319	0.5186; 0.6820	0.6869; 0.7473	0.5186; 0.6820	0.5179; 0.6835
70:30	Before; After	0.6222; 0.6453	0.6159; 0.7112	0.7527; 0.8202	0.6159; 0.5400	0.6163; 0.7121
80:20	Before; After	0.6575; 0.6540	0.6540; 0.6268	0.8085; 0.8849	0.6540; 0.7178	0.6541; 0.5859
90:10	Before; After	0.6841; 0.6610	0.6892; 0.6402	0.8590; 0.9426	0.6892; 0.8126	0.6932; 0.6813
95:1	Before; After	0.6896; 0.6732	0.7053; 0.6732	0.9028; 0.9130	0.8926; 0.7608	0.7076; 0.6733
99:1	Before; After	0.7574; 0.6829	0.7212; 0.7373	0.8328; 0.7143	0.7212; 0.7538	0.7296; 0.6775
AVG	Before; After	0.6456; 0.6516	0.6265; 0.6735	0.7307; 0.8124	0.6532; 0.7016	0.6284; 0.6661