Analysis of Changes in Market Shares of Commercial Banks Operating in Turkey Using Computational Intelligence Algorithms

Table 8

The rules generated by the best algorithm for each year (dataset).

Year	Algorithm ID	Rule
1990	10	MS = 0.97 * GS3 − 0.02
1991	10	MS = 0.99 * GS3 − 0.02
1992	11	MS = −0.0292 * (normalized) L2; + 0.3947 * (normalized) GS2; + 0.5996 * (normalized) GS3; + 0.0068
1993	11	MS = −0.1056 * (normalized) L2; + 0.0094 * (normalized) P3; + 0.0052 * (normalized) IE2; + 0.4708 * (normalized) GS1; + 0.523 * (normalized) GS3; + 0.0044
1994	10	MS = 1.04 * GS1 − 0.02
1995	10	MS = 0.99 * GS1 − 0.02
1996	10	MS = 1 * GS1 − 0.01
1997	10	MS = 0.97 * GS3 + 0
1998	10	MS = 1.06 * GS3 + 0.01
1999	11	MS = −0.0237 * (normalized) L1; − 0.0225 * (normalized) IE2; + 0.296 * (normalized) GS1; + 0.3055 * (normalized) GS2; + 0.3722 * (normalized) GS3; + 0.0272
2000	8	MS = −0.1229 * L1 + 0.9571 * GS3 + 0.0769
2001	3	If GS1 in (−inf–0.1] then MS = 0.03137; If GS1 in (0.3–0.4] then MS = 0.59189; If GS1 in (0.4–0.5] then MS = 0.041797; If GS1 in (0.5–0.6] then MS = 0.441048; If GS1 in (0.8–0.9] then MS = 0.093575; If GS1 in (0.9–inf) then MS = 1.0
2002	15, 11	The best performed algorithm is bagging. But the rules generated by bagging are very long. Instead of giving these long rules, a rule generated by the second best algorithm (SMOreg) is given.; MS = −0.2014 * (normalized) CA2; − 0.0089 * (normalized) L1; − 0.0859 * (normalized) IE2; + 0.7895 * (normalized) GS1; + 0.1961
2003	11	MS = −0.0024 * (normalized) L1; + 0.0009 * (normalized) P2; + 0.3401 * (normalized) GS1; + 0.2689 * (normalized) GS2; + 0.4899 * (normalized) GS3; − 0.0007
2004	11	MS = −0.0015 * (normalized) P2; + 0.342 * (normalized) GS1; + 0.2547 * (normalized) GS2; + 0.4802 * (normalized) GS3; − 0.0001
2005	11	MS = −0.0028 * (normalized) IE3; + 0.3304 * (normalized) GS1; + 0.3268 * (normalized) GS2; + 0.4712 * (normalized) GS3; + 0.0002
2006	8	MS = 0.1241 * GS1 + 0.3862 * GS2 + 0.6142 * GS3 + −0.0026
2007	11	MS = 0.0049 * (normalized) CA3; + 0.0012 * (normalized) P2; + 0.3388 * (normalized) GS1; + 0.3281 * (normalized) GS2; + 0.471 * (normalized) GS3; − 0.0056
2008	3	If L3 in (−inf–0.1] then MS = 0.1942307; If L3 in (0.1–0.2] then MS = 0.479808; If L3 in (0.2–0.3] then MS = 0.073077; If L3 in (0.3–0.4] then MS = 0.019231; If L3 in (0.4–0.5] then MS = 0.980769; If L3 in (0.5–0.6] then MS = 0.2230765; If L3 in (0.6–0.7] then MS = 0.0826925; If L3 in (0.7–0.8] then MS = 0.042308; If L3 in (0.9–inf) then MS = 0.242308
2009	1	MS = 0.403535