快速的方法来计算滑动窗口上的事件 [英] Fast way to count events over a sliding window

问题描述

假设我有 x = rnorm（100000），而不是做一个 1000 长度滑动窗口移动平均，我想要做一个 1000 长度滑动窗口，它计算 x 高于 0.2 。

例如，

  x < -  rnorm 
 start<  -  1：1000 
 record<  -  list（）
 while（start [length（start）]< = length（x））{
 record [ [length（record）+ 1]] < -  length（which（x [start]> 0.2））/ length（start）
 start < -  start + 1 
 print length（record）]]）; flush.console（）
} 
  

对于大 length（x）。什么是高效的方法？

解决方案

我的贡献是计算条件的累积和之间的滞后差>

  cumdiff = function（x）diff（c（0，cumsum（x> .2）），20）
  

以及

  filt = function（x）filter（x> 0.2，rep（1,20），sides = 1）
 library（TTR）; ttr = function（x）runSum（x> .2,20）
 cumsub = function（x）{z- cumsum（c（0，x> 0.2））; tail（z，-20） -  head（z，-20）} 
  

>

 >库（microbenchmark）
> set.seed（123）; xx = rnorm（100000）
> microbenchmark（cumdiff（xx），filt（xx），ttr（xx），cumsub（xx））
单位：毫秒
 expr min lq median uq max neval 
 cumdiff（xx）11.192005 12.387862 12.469253 12.77588 13.72404 100 
 filt（xx）20.979503 22.058045 22.442765 23.02612 62.91730 100 
 ttr（xx）8.390923 10.023934 10.119772 10.46309 11.04173 100 
 cumsub（xx）7.015654 8.483432 8.538171 8.73596 9.65421 100 
  

这些不同在于如何表示结果的细节（ filt 和 ttr 具有领先的NAs），并且只有过滤器处理嵌入式NA的

 > xx [22] = NA 
> head（cumdiff（xx））＃NA's propagate，silently 
 [1] 9 9 NA NA NA NA 
> ttr（xx）
在runSum中的错误（x> 0.2，20）：系列包含非领先的NAs 
>尾（filt（xx），-19）
 [1] 9 9 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA 8 8 9 
 ... 
  

Suppose I have x = rnorm(100000) and instead of doing a 1000 length sliding window moving average, I wanted to do a 1000 length sliding window that counts all the times that x is above 0.2.

For example,

x <- rnorm(1004)
start <- 1:1000
record <- list()
while(start[length(start)] <= length(x)) {
    record[[length(record) + 1]] <- length(which(x[start] > 0.2))/length(start)
    start <- start + 1
    print(record[[length(record)]]);flush.console()
}

This becomes unmanegable for large length(x). What is a highly efficient method?

解决方案

My contribution is to calculate the lagged difference between the cumulative sum of the condition

cumdiff = function(x) diff(c(0, cumsum( x > .2)), 20)

which along with

filt = function(x) filter(x > 0.2, rep(1, 20), sides=1)
library(TTR); ttr = function(x) runSum(x > .2, 20)
cumsub = function(x) { z <- cumsum(c(0, x>0.2)); tail(z,-20) - head(z,-20) }

performs ok

> library(microbenchmark)
> set.seed(123); xx = rnorm(100000)
> microbenchmark(cumdiff(xx), filt(xx), ttr(xx), cumsub(xx))
Unit: milliseconds
        expr       min        lq    median       uq      max neval
 cumdiff(xx) 11.192005 12.387862 12.469253 12.77588 13.72404   100
    filt(xx) 20.979503 22.058045 22.442765 23.02612 62.91730   100
     ttr(xx)  8.390923 10.023934 10.119772 10.46309 11.04173   100
  cumsub(xx)  7.015654  8.483432  8.538171  8.73596  9.65421   100

These differ in the specifics of how the result is represented (filt and ttr have leading NAs, for instance) and only filter deals with embedded NA's

> xx[22] = NA
> head(cumdiff(xx))  # NA's propagate, silently
[1]  9  9 NA NA NA NA
> ttr(xx)
Error in runSum(x > 0.2, 20) : Series contains non-leading NAs
> tail(filt(xx), -19)
 [1]  9  9 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA  8  8  9
 ...

这篇关于快速的方法来计算滑动窗口上的事件的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！