寻找两个图之间的相似性度量..？
抱歉，如果我的问题听起来很业余..
实际上我有一组二维形式的图
让 X=(x1,x2...xn) 是一组获得的类似图
Y=(y1,y2...yn) be a set of plots similar
直观地我可以看到 X 的所有图“看起来相似”但是我如何找到两个图之间的分数之间的相似性并证明它们具有很高的相似性分数..??
我正在使用 R 语言...有人可以帮忙吗...??谢谢
sorry if my question sounds very amateurish..
Actually I have a set of plots in 2d form
Let X=(x1,x2...xn) be a set of similar plots obtained
Y=(y1,y2...yn) be a set of plots similar
Intuitively i can see that all plots of X 'look similar' But how do i find the similarity between scores between 2 plots and prove that they have a high similarity score..??
I am using the R language... Can somebody help..??Thanks
如果你对这篇内容有疑问，欢迎到本站社区发帖提问 参与讨论，获取更多帮助，或者扫码二维码加入 Web 技术交流群。
绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱，如果其他用户或者作者回复了您的评论，将不能在第一时间通知您！
发布评论
评论（2）
您似乎正在考虑两个图在任何给定位置都有值的情况。也许这个方法会起作用：
对于每个索引 i，计算 (xi-yi)^2。
所有 i 的总和。
除以 n。
这只是计算图中各点之间的平均差异，因此 0 表示完全相同，而较大的值意味着相似度较低。从统计学上来说，可能有一种更准确的方法，但这肯定是一个很好的估计。
It seems that you are thinking of the case in which both plots have a value at any given position. Maybe this method will work:
For each index i, calculate (xi-yi)^2.
Sum over all i.
Divide by n.
This just calculates the average difference between points in the plot, so 0 would be exactly the same, while larger values mean less similarity. Statistically, there's probably a more accurately method, but this is a good estimate for sure.
看来我可能有点晚了，但是 FastDTW 是对动态时间扭曲并实施here in python 听起来正是您所需要的！
Seems I may be a bit late here, but FastDTW, an improvement on Dynamic Time Warping and implemented here in python sounds like exactly what you need!