|
另外两个审稿人的审稿意见都是条件接收,但其中一个审稿人要求增加额外的实验,原话如下:
(1) Additional experiments.
To conclude that the proposed method (SC-FD) outperforms the method in [2], the authors should show comparison on more than one dataset.
我在论文中只用了一个数据集做实验,那我是不是再增加一个数据集就可以了,还是要增加两个?
我用公开数据集做实验时因为论文的目的只用了一部分数据,所以没办法跟其他论文的精度做比较,但我会建立一个基线算法如卷积神经网络做比较,这样可以吗,谢谢各位啦
Reviewer A
(1) Additional experiments.
To conclude that the proposed method (SC-FD) outperforms the method in [2], the authors should show comparison on more than one dataset. For example, reporting results on Skoda, Daphne FoG and/or WISDM v2.0 as in [2] would be good choices.
(2) Analysis of hidden relations.
------------
Reviewer B
-----------
1. The equations expressed in the paper must be thoroughly reviewed.
2. The CNN models presented in the paper do not seem to include any fully-connected layer.
3. In p. 4, top right column, the first sentence reads, "the dashed line in Figure 2." There is no dashed line in Figure 2.
4. In p. 8 left column, last sentence of the first paragraph reads, "SC-BA can exploit temporal correlation of sensor [13]."
I believe it is 'spatial', not 'temporal'.
5. The paper is clearly written overall, but could be improved. |
|