Table11

Table 11 COMPARISON OF GENETIC ADAPTIVE NEURAL NETWORK (GANN) APPROXIMATIONS OF EURODOLLAR FUTURES OPTION TRADED ON LIFFE Panel A: Calls
		GANN₁ Approximation		GANN₂ Approximation
Sub-sample Description	Number of Observations	Mean Squared Error (MSE)	Mean Absolute Error (MAE)	Mean Squared Error (MSE)	Mean Absolute Error (MAE)
Complete Sample	6887	0.00051	0.01677	0.00021	0.01018
in-the-money	2698	0.00071	0.02098	0.00028	0.01197
just-in	1782	0.00076	0.02159	0.00036	0.01380
deep-in	910	0.00061	0.01978	0.00014	0.00838
out-of-the-money	4159	0.00038	0.01399	0.00016	0.00898
just-out	1840	0.00059	0.01722	0.00029	0.01251
deep-out	2312	0.00022	0.01143	0.00006	0.00617
at-the-money	30	0.00117	0.02360	0.00043	0.01500
Panel B: Puts
Complete Sample	6887	0.00046	0.01632	0.00020	0.01014
in-the-money	4159	0.00041	0.01511	0.00016	0.00875
just-in	1840	0.00056	0.01740	0.00029	0.01261
deep-in	2312	0.00029	0.01328	0.00006	0.00566
out-of-the-money	2698	0.00054	0.01806	0.00027	0.01225
just-out	1782	0.00062	0.01898	0.00034	0.01368
deep-out	910	0.00039	0.01627	0.00014	0.00948
at-the-money	30	0.00118	0.02606	0.00036	0.01308
The degree of moneyness is as previously described. For example, if the futures rate (F(t)) is greater than the strike rate (100 - K), a put option on a 3-month Eurodollar futures contract will be in-the-money and a call option will be out-of-the-money. GANN₁ has 4 inputs and 18 hidden layer nodes while GANN₂ has 6 input and 18 hidden layer nodes.