We use cookies to ensure you have the best browsing experience on our website. Please read our cookie policy for more information about how we use cookies.
  1. Prepare
  2. Tutorials
  3. 10 Days of Statistics
  4. Day 7: Pearson Correlation Coefficient I
  5. Discussions

Day 7: Pearson Correlation Coefficient I

Sort by

recency

|

104 Discussions

|

  • JanekPython09
     + 0 comments

    python sol

    import math 

n = int(input())
x = list(map(float, input().split()))
y = list(map(float, input().split()))
lx = len(x)
ly = len(y)

meanX = sum(x)/lx
meanY = sum(y)/ly

sdX = math.sqrt(sum([math.pow(xi-meanX, 2) for xi in x])/lx)
sdY = math.sqrt(sum([math.pow(yi-meanY, 2) for yi in y])/ly)


PCC = (sum((xi-meanX)*(yi-meanY) for xi, yi in zip(x, y)))/(n*sdX*sdY)

print(round(PCC, 3))

  • akshaymoolath
     + 0 comments
    #python
from statistics import pstdev, mean
n=int(input())
X=list(map(float, input().split()))
Y=list(map(float, input().split()))
res=[(X[i]-mean(X))*(Y[i]-mean(Y)) for i in range(n)]
p=sum(res)/(n*pstdev(X)*pstdev(Y))
print(round(p, 3))

  • razawy_h
     + 0 comments

    I am bit confused why its not mentioned to take the population standard deviation instead of the sample standarddeviation 

    # Enter your code here. Read input from STDIN. Print output to STDOUT
import math
import statistics
n = int(input())


X =list(map(float, input().split()))


Y =list(map(float, input().split()))
    

def calcpearsonCov(dataset1, dataset2):
    averagex= statistics.mean(dataset1)
    averagey = statistics.mean(dataset2)
    result = 0
    for i in range(0, len(dataset1)):
        result += (dataset1[i] - averagex) * (dataset2[i]- averagey)
    #result = result *(1/len(dataset1))
    stdevx = statistics.pstdev(dataset1)
    stdevy = statistics.pstdev(dataset2)
    result = result/(n*(stdevx*stdevy))
    return result
    
result = calcpearsonCov(X, Y)
print(round(result, 3))

  • sylvain_michel
     + 0 comments

    in R could be 

    stdin <- file('stdin')
open(stdin)
n <- as.integer(trimws(readLines(stdin, n = 1, warn = FALSE), which = "both"))
dataX <- as.numeric(strsplit(trimws(readLines(stdin, n = 1, warn = FALSE), which = "right"), " ")[[1]])
dataY <- as.numeric(strsplit(trimws(readLines(stdin, n = 1, warn = FALSE), which = "right"), " ")[[1]])
correlation = cor(dataX, dataY, method = 'pearson')
cat(round(correlation, 3))

  • kushalkharelsea1
     + 0 comments

    Easy Solution:

    # import statistics as stat
from sys import stdin, stdout
import math

def mean(arr):
    return sum(arr) / len(arr)

def sd(arr, mean):
    squared_diff_sum = sum((x - mean)**2 for x in arr)
    variance = squared_diff_sum / len(arr)
    return math.sqrt(variance)

def covariance(X, Y):
    mean_X = mean(X)
    mean_Y = mean(Y)
    covariance_sum = sum((X[i] - mean_X) * (Y[i] - mean_Y) for i in range(len(X)))
    cov = covariance_sum / len(X)
    return cov

def pearson_correlation(X, Y):
    cov_XY = covariance(X, Y)
    st_dev_X = sd(X, mean(X))
    st_dev_Y = sd(Y, mean(Y))
    coefficient = cov_XY / (st_dev_X * st_dev_Y)
    return coefficient

n = int(stdin.readline().strip())
X = list(map(float, stdin.readline().strip().split()))
Y = list(map(float, stdin.readline().strip().split()))

if len(X) != len(Y):
    print("Error: Data sets X and Y must have equal lengths")
else:
    t = pearson_correlation(X, Y)
    print(round(t, 3))