Exhaustive Search and Power-Based Gradient Descent Algorithms for Time-Delayed FIR Models

Table 1

The parameter estimates and their estimation errors.

Algorithms
T-GD	1	−0.00076	0.00014	0.04923	0.06527	−0.03305	0.02249	0.07435	0.00028	−0.00061	0.00039	90.22459
	2	−0.00139	0.00027	0.09359	0.12417	−0.06288	0.04280	0.14142	0.00050	−0.00111	0.00070	81.40546
	5	−0.00268	0.00062	0.20209	0.26859	−0.13612	0.09259	0.30580	0.00092	−0.00205	0.00134	59.79451
	10	−0.00351	0.00104	0.32227	0.42938	−0.21783	0.14806	0.48858	0.00109	−0.00248	0.00172	35.76318
	15	−0.00354	0.00132	0.39373	0.52564	−0.26688	0.18129	0.59784	0.00096	−0.00227	0.00171	21.39891
	20	−0.00328	0.00150	0.43622	0.58328	−0.29632	0.20121	0.66316	0.00073	−0.00186	0.00157	12.81303
ES-GD	1	0.01755	−0.02912	0.31694	0.47762	−0.26669	0.15867	0.52597	−0.04381	0.01115	0.02908	31.10412
	2	0.01217	−0.01885	0.43294	0.61302	−0.32675	0.20761	0.68377	−0.02656	0.00853	0.01859	10.47646
	5	0.00153	−0.00351	0.49623	0.66886	−0.34469	0.22932	0.75506	−0.00126	0.00332	−0.00035	0.78800
	10	0.00023	−0.00230	0.49948	0.67068	−0.34518	0.23005	0.75812	0.00101	0.00266	−0.00251	0.59350
	15	0.00021	−0.00229	0.49951	0.67069	−0.34519	0.23005	0.75815	0.00103	0.00265	−0.00254	0.59388
	20	0.00021	−0.00229	0.49951	0.67069	−0.34519	0.23005	0.75815	0.00103	0.00265	−0.00254	0.59389
P-GD	1	0.02557	−0.04243	0.46184	0.69598	−0.38862	0.23121	0.76643	−0.06384	0.01625	0.04237	9.53122
	2	0.00246	−0.00123	0.49699	0.66530	−0.33847	0.22943	0.75110	0.00198	0.00327	0.00074	0.96875
	5	0.00022	−0.00229	0.49950	0.67070	−0.34520	0.23005	0.75814	0.00102	0.00266	−0.00253	0.59415
	10	0.00021	−0.00229	0.49951	0.67069	−0.34519	0.23005	0.75815	0.00103	0.00265	−0.00254	0.59389
	15	0.00021	−0.00229	0.49951	0.67069	−0.34519	0.23005	0.75815	0.00103	0.00265	−0.00254	0.59389
	20	0.00021	−0.00229	0.49951	0.67069	−0.34519	0.23005	0.75815	0.00103	0.00265	−0.00254	0.59389
True values		0.00000	0.00000	0.50000	0.67000	−0.34000	0.23000	0.76000	0.00000	0.00000	0.00000	0.00000