Question about r value in linear regression.

[CodeWarrior14]

I'm writing a program tha calculates linear regression, one of the values that are generated is the r (linear correlation coefficient) value. However, my algorithm doesnt work.

I implemented the r value from the formula located at Click Here

Here is my code:

long double LinearRegression::calculate_r()
{
    long double r;
    long double sumxy = 0;
    long double sumx = 0;
    long double sumy = 0;
    long double numerator;
    long double denominator;

    long double sumx2 = 0;
    long double sumxto2 = 0;
    long double sumy2 = 0;
    long double sumyto2 = 0;
    long double sqrt1 = 0;
    long double sqrt2 = 0;

    int arrIt;

    // Calcualte the sumation of x, y and x * y.
    for (arrIt = 0; arrIt < this->n; arrIt++)
    {
        sumxy += this->x_list[arrIt] * this->y_list[arrIt];
        sumx += this->x_list[arrIt];
        sumy += this->y_list[arrIt];
    }

    numerator = this->n * sumxy - sumx * sumy;

    for (arrIt = 0; arrIt < this->n; arrIt++)
    {
        sumx2 += pow(this->x_list[arrIt], 2);
        sumy2 += pow(this->y_list[arrIt], 2);
    }

    sumxto2 = pow(sumx2, 2);
    sumyto2 = pow(sumy2, 2);

    sqrt1 = sqrt(this->n * sumx2 - sumxto2);
    sqrt2 = sqrt(this->n * sumy2 - sumyto2);

    denominator = sqrt1 * sqrt2;

    r = numerator / denominator;

    this->r = r;

    return r;
}

I get a value of: -1.#IND
I am suppose to get a value of: 0.999997556

Whats wrong with my implementation and how do I fix it?

[mike_2000_17]

I believe that the problem is that your lines 35-36:

sumxto2 = pow(sumx2, 2);
sumyto2 = pow(sumy2, 2);

should be as follows:

sumxto2 = pow(sumx, 2);
sumyto2 = pow(sumy, 2);

note the use of sumx and sumy instead of the sumx2 and sumy2.

[CodeWarrior14]

That fixed it. Thanks