gpt4 book ai didi

Why Rolling Wavelet (Using Pywavelet Package) Is Not Casual?(为什么滚动小波(使用PYWORWEL软件包)不是随机的?)

转载 作者:bug小助手 更新时间:2023-10-28 13:41:31 25 4
gpt4 key购买 nike



summarily and Clearly Speaking, There is a Time series Data Set (Suppose 1000 Sample), I take 100 first sample (0 to 99) and get wavelet and reconstruct it (or equivalently, denoise it), Next Shift the window by 1 (1 to 100) and do so again, Just Like Move Mean Algorithm.
But There is an issue, I see high fluctuation in the signal edge (see red rectangle in below figure).
**What is the cause of this phenomena?
**# Is This a regular behavior?

概括地说,有一个时间序列数据集(假设1000个样本),我取100个第一个样本(0到99),得到小波并重建它(或等效地,去噪它),然后将窗口移动1(1到100),再做一次,就像移动均值算法一样。但有一个问题,我看到信号边缘的高波动(见下图中的红色矩形)。** 这种现象的原因是什么?**#这是一个正常的行为吗?


Or What is wrong?**


enter image description here

在此处输入图像描述


How can I get rolling wavelet?

怎样才能得到滚动子波?


更多回答

Welcome to SO @mohammad. Could you please share a code example with us? You may edit your original post to include your script.

欢迎来到如此“穆罕默德”。您能和我们分享一个代码示例吗?您可以编辑您的原始帖子以包括您的脚本。

Without seeing the code, my best guess is that it's because there are not enough signal points left for all filter coefficients (e.g. when you have a filter length of 10 and are 3 positions away from the end). A common solution is to wrap around, so after signal point x[n-1] go back to x[0]. As you can see, in your example x[0] is much higher than x[n-1] so that may explain the higher signal in your reconstruction. In other words, you do have e 'rolling' wavelet, if that means going from n-1 back to 0. There is Not Much Clarity without showing the code though.

在没有看到代码的情况下,我最好的猜测是,这是因为没有足够的信号点用于所有的滤波器系数(例如,当您的滤波器长度为10并且距离末尾有3个位置时)。一种常见的解决方案是回绕,因此在信号点x[n-1]之后返回到x[0]。如您所见,在您的示例中,x[0]比x[n-1]高得多,因此这可能解释了重建过程中信号较高的原因。换句话说,如果这意味着从n-1回到0,你确实有e个‘滚动’子波。不过,如果不显示代码,就不会有太多的清晰度。

优秀答案推荐
更多回答

25 4 0
Copyright 2021 - 2024 cfsdn All Rights Reserved 蜀ICP备2022000587号
广告合作:1813099741@qq.com 6ren.com