Pearson Correlation (JavaScript)

pearson-correlation.js

/**
 *  @fileoverview Pearson correlation score algorithm.
 *  @author matt.west@kojilabs.com (Matt West)
 *  @license Copyright 2013 Matt West.
 *  Licensed under MIT (http://opensource.org/licenses/MIT).
 */


/**
 *  Calculate the person correlation score between two items in a dataset.
 *
 *  @param  {object}  prefs The dataset containing data about both items that
 *                    are being compared.
 *  @param  {string}  p1 Item one for comparison.
 *  @param  {string}  p2 Item two for comparison.
 *  @return {float}  The pearson correlation score.
 */
function pearsonCorrelation(prefs, p1, p2) {
  var si = [];

  for (var key in prefs[p1]) {
    if (prefs[p2][key]) si.push(key);
  }

  var n = si.length;

  if (n == 0) return 0;

  var sum1 = 0;
  for (var i = 0; i < si.length; i++) sum1 += prefs[p1][si[i]];

  var sum2 = 0;
  for (var i = 0; i < si.length; i++) sum2 += prefs[p2][si[i]];

  var sum1Sq = 0;
  for (var i = 0; i < si.length; i++) {
    sum1Sq += Math.pow(prefs[p1][si[i]], 2);
  }

  var sum2Sq = 0;
  for (var i = 0; i < si.length; i++) {
    sum2Sq += Math.pow(prefs[p2][si[i]], 2);
  }

  var pSum = 0;
  for (var i = 0; i < si.length; i++) {
    pSum += prefs[p1][si[i]] * prefs[p2][si[i]];
  }

  var num = pSum - (sum1 * sum2 / n);
  var den = Math.sqrt((sum1Sq - Math.pow(sum1, 2) / n) *
      (sum2Sq - Math.pow(sum2, 2) / n));

  if (den == 0) return 0;

  return num / den;
}

The above code has a bug, n is used before being defined.

let n = x.length

needs to be before the const calc ... line.

The above code has a bug, n is used before being defined.

let n = x.length

needs to be before the const calc ... line.

No it hasn't. The third line defines function calc without calling it.

The above code has a bug, n is used before being defined.
let n = x.length
needs to be before the const calc ... line.

No it hasn't. The third line defines function calc without calling it.

Well, it's still bad form. You're referencing a global that's defined later in scope and generally that's bad form.
The current version of ESLint doesn't like this, but as you're only giving an example here I guess it might be fine.