pine-script - 我的Spearman相关系数脚本无法正常工作(My Spearman Correlation Coefficient script is not working correctly)

Question

I'm trying to do a Spearman rank correlation coefficient script in Pine Script.

(我正在尝试在Pine Script中制作Spearman等级相关系数脚本。)

The problem i have is it sometimes spikes down to two digit negative values.

(我的问题是，有时它会下降到两位数的负值。)

This script is a compression to the built in correlation() (Pearson Coefficient) function and my Spearman behaves differently.

(该脚本是对内置的correlation（）（皮尔逊系数）函数的压缩，我的斯皮尔曼的行为有所不同。)

So I think it has to be something wrong to my math or is this the expected behavior Spearman correlation?

(所以我认为我的数学一定有问题，还是这是Spearman相关的预期行为？)

Here is the script:

(这是脚本：)

//@version=4
study("Spearman Rank Correlation Coefficient [ChuckBanger]", shorttitle="Spearman Correlation CB", precision=3)

length = input(title="Length", type=input.integer, defval=20)
sym1 = input(title="1st Symbol", type=input.symbol, defval="BITMEX:XBTUSD")
sym2 = input(title="2nd Symbol", type=input.symbol, defval="BITMEX:ETHUSD")
src = input(title="Source", type=input.source, defval=close)

data1 = security(sym1, timeframe.period, src)
data2 = security(sym2, timeframe.period, src)

rankit(x, n) =>
    int r=0, int s=0
    float rank = na
    for i = 0 to n by 1
        r:=1,s:=1
        for j = 0 to i by 1
            if (x[j] < x[i])
                r := r + 1
            if (x[j] == x[i]) 
                s := s + 1

        for k = i + 1 to n by 1
            if (x[k] < x[i])
                r := r + 1
            if (x[k] == x[i]) 
                s := s + 1

        rank := r + (s - 1) * 0.5

correlationCoefficient(X, Y, n) =>
    float sum_X = 0
    float sum_Y = 0
    float sum_XY = 0
    float squareSum_X = 0
    float squareSum_Y = 0
    for i = 0 to n by 1 
        // sum of elements of array X. 
        sum_X := nz(sum_X) + X[i]

        // sum of elements of array Y. 
        sum_Y := nz(sum_Y) + Y[i]

        // sum of X[i] * Y[i]. 
        sum_XY := nz(sum_XY) + X[i] * Y[i]

        // sum of square of array elements. 
        squareSum_X := squareSum_X + X[i] * X[i]
        squareSum_Y := squareSum_Y + Y[i] * Y[i]

    corr = (n * sum_XY - sum_X * sum_Y) / sqrt((n * 
     squareSum_X - sum_X * sum_X) * (n * squareSum_Y - sum_Y * sum_Y))
    corr


spearman(data1, data2, length) =>
    rank_x = rankit(data1, length)
    rank_y = rankit(data2, length)
    correlationCoefficient(rank_x, rank_y, length)

spear = spearman(data1, data2, length)
spearCorr = sym1 == sym2 ? 1 : spear
pearsonCorr = correlation(data1, data2, length)

plot(spearCorr, title="Spearman", color=color.orange, transp=0)
plot(pearsonCorr, title="Pearson", color=color.purple, transp=0)

  ask by ChuckBanger translate from so

Answer 1

等待大神答复