|
Fig. 2. The result of the model performance changing with the forecast target in SSEC data set. Table 7An adaptive feature selection schema using improved technical indicatorsFig. 2. The result of the model performance changing with the forecast target in SSEC data set.
Table 7
The number of features before and after feature selection of size-varied time windows in four data sets. Time window Feature number Feature number after feature selection 1
SSEC
HSI
DJI SP 500 0
18 18 18 18 18 3
72 30 43 35 42 5
108 42 37 39 30 10 198 25 31 34 32 15 288 28 35 55 51 30 558 42 44 51 55 45 828 38 32 69 63 60 1098 53 44 65 49
Table 8
SSEC data set feature subset and the corresponding model performance. Feature number in the feature subsets Accuracy Precision Recall F score
0 18 0.725 0.729 0.725 0.724 1
17 0.724 0.725 0.725 0.725 2
16 0.725 0.729 0.725 0.724 3
15 0.725 0.730 0.725 0.725 4
14 0.727 0.730 0.729 0.729 5
13 0.730 0.730 0.730 0.730 6
12 0.728 0.729 0.726 0.726 7
11 0.729 0.732 0.730 0.730 8
10
0.733
0.732
0.732 0.732
9 9
0.728 0.729 0.726 0.726 10 8
0.729 0.732 0.730 0.730 11 7
0.730 0.730 0.730 0.730 12 6
0.730 0.730 0.730 0.730 13 5
0.727 0.728 0.726 0.726 14 4
0.719 0.719 0.718 0.718 15 3
0.707 0.706 0.706 0.706 16 2
0.683 0.683 0.683 0.683 17 1
0.627 0.630 0.630 0.630
Table 9
HSI data set feature subset and the corresponding model performance. Feature number in the feature subsets Accuracy Precision Recall F score
0 18 0.763 0.764 0.764 0.764 1
17 0.766 0.766 0.766 0.766 2
16 0.769 0.769 0.769 0.769 3
15 0.766 0.768 0.768 0.768 4
14 0.772 0.771 0.771 0.771 5
13 0.775 0.774 0.774 0.774 6
12 0.780 0.781 0.781 0.781 7
11 0.784 0.784 0.784 0.784 8
10 0.779 0.780 0.780 0.780 9
9 0.783 0.782 0.782 0.782 10 8
0.777 0.776 0.776 0.776 11 7
0.778 0.779 0.779 0.779 12 6
0.786
0.785
0.785 0.785
13 5
0.769 0.770 0.770 0.770 14 4
0.767 0.768 0.768 0.768 15 3
0.751 0.751 0.751 0.751 16 2
0.728 0.729 0.729 0.729 17 1
0.661 0.660 0.660 0.660
Table 10
DJI data set feature subset and the corresponding model performance. Feature number in the feature subsets Accuracy Precision Recall F score
0 18 0.721 0.729 0.722 0.703 1
17 0.717 0.734 0.716 0.696 2
16 0.730 0.738 0.729 0.715 3
15 0.741 0.745 0.741 0.732 4
14 0.735 0.745 0.736 0.720 5
13 0.728 0.740 0.727 0.710 6
12 0.728 0.739 0.728 0.712 7
11 0.728 0.739 0.729 0.714 8
10 0.724 0.733 0.726 0.710 9
9 0.737 0.738 0.738 0.728 10 8
Share with your friends: |
|
|
