Open Access

表1

不同知识图谱构建方法对检索与分类性能的影响

数据集 损失函数 方法 准确率/% 检索性能
mAP↑ FT↑ ST↑ NN↑ DCG↑ ANMRR↓
注:表 1~4中黑体数字表示该列一个数据集下最好的性能, “↑”表示数值越大时性能越好,“↓”反之
FI L1 64.98 0.651 7 0.660 4 0.700 6 0.661 5 0.798 6 0.383 7
65.65 0.686 2 0.686 7 0.733 0 0.663 6 0.803 2 0.288 3
66.29 0.697 4 0.695 7 0.751 8 0.681 1 0.811 2 0.256 4
Ldot 65.41 0.654 4 0.677 0 0.738 1 0.662 2 0.796 5 0.375 1
Emotion6 L1 56.11 0.488 5 0.593 5 0.605 8 0.488 4 0.784 5 0.420 8
56.60 0.506 1 0.606 8 0.622 5 0.493 1 0.789 0 0.379 6
Ldot 55.26 0.487 2 0.584 4 0.607 6 0.476 3 0.779 4 0.510 4

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.