A Stock Selection Model of Image Classification Method Based on Convolutional Neural Network

Table 3

Results of FNN under different parameters.

Training size	Hidden layer	Benchmark	Annual return	Sharp ratio	Max drawdown
12	1	64367	5237	0.0023	0.2080
24	1	29864	1408	0.0192	0.2203
48	1	35711	11342	0.0043	0.2309
12	2	64367	32881	0.0472	0.2114
24	2	29864	4636	0.0150	0.2207
48	2	35711	27498	0.0236	0.2348
12	3	64367	38002	0.0352	0.2114
24	3	29864	26367	0.0250	0.1913
48	3	35711	29701	0.0236	0.1967