Predicting Subcontractor Performance Using Web-Based Evolutionary Fuzzy Neural Networks

Table 1

Subcontractor performance historical records.

Pattern no.	Performance	Input			Output
		Last 3	Last 2	Last 1	Normalized performance
Input patterns
1	80	72	76	76	0.8333
2	86	76	76	80	0.8958
3	74	80	76	76	0.7708
4	70	76	76	74	0.7292
5	68	56	62	66	0.7083
6	66	60	66	68	0.6875
7	66	70	72	68	0.6875
8	58	62	66	60	0.6042
9	56	66	60	58	0.5833
10	80	76	74	76	0.8333
11	86	74	76	80	0.8958
12	88	76	80	86	0.9167
13	76	86	80	80	0.7917
14	70	66	68	66	0.7292
15	70	68	66	70	0.7292
16	76	66	70	70	0.7917
17	74	66	70	76	0.7708
18	76	70	76	74	0.7917
19	80	74	76	76	0.8333
20	66	62	58	62	0.6875
21	68	58	62	66	0.7083
22	76	76	74	76	0.7919
Test patterns
23	66	62	56	60	0.6875
24	68	56	60	66	0.7083
25	66	60	66	68	0.6875
26	66	66	68	66	0.6875
27	70	68	66	66	0.7292
28	76	66	66	70	0.7917
29	74	66	70	76	0.7708
30	76	70	76	74	0.7917
31	76	76	74	76	0.7917
32	80	74	76	76	0.8333
33	86	76	76	80	0.8958
34	88	76	80	86	0.9167

Note: Last 1 denotes the subcontractor’s latest performance, and so forth. Normalized performance is divided by 96.