为什么以下 Javascript 代码的输出是 NaN？

Question

我是 JavaScript 的新手。我正在尝试编写一个代码，通过对一组非常小的数据使用 Pearson 相关系数公式来查找相关系数。这是皮尔逊相关系数的公式。

var dataSet = {
  bath: [30, 29, 20, 28, 12, 30],
  disease: [2, 5, 9, 3, 8]
};
const corRelate = () => {
  var m1 = 0,
    m2 = 0,
    m3 = 0,
    m4 = 0,
    m5 = 0;
  for (var i = 0; i < dataSet.bath.length; i++) {
    m2 = m2 + dataSet.bath[i];
    m4 = m4 + dataSet.bath[i] * dataSet.bath[i];
  }

  for (var i = 0; i < dataSet.disease.length; i++) {
    m3 = m3 + dataSet.disease[i];
    m5 = m5 + dataSet.disease[i] * dataSet.disease[i];
  }

  for (var i = 0; i < dataSet.bath.length; i++) {
    for (var j = 0; j < dataSet.disease.length; j++) {
      m1 = m1 + dataSet.bath[i] * dataSet.disease[j];
    }
  }
  const val = (m1 - m2 * m3) / (Math.sqrt(m4 - m2 * m2) * Math.sqrt(m5 - m3 * m3));
  return val;
};

const val = corRelate();
console.log(val);

这段代码的输出是 NaN 。我认为计算的输出应该是整数或浮点数，我不明白为什么我得到 NaN 作为输出。

Answer 1

你的公式有误，我重写了你的代码片段：

const dataSet = {
  bath: [30, 29, 20, 28, 12, 30],
  disease: [2, 5, 9, 3, 8]
};

const corRelate = () => {
  const m1 = dataSet.bath.reduce((m1, bath) => {
    return dataSet.disease.reduce((m1, disease) => m1 + bath * disease, m1);
  }, 0);

  const m2 = dataSet.bath.reduce((m2, item) => m2 + item, 0);
  const m4 = dataSet.bath.reduce((m4, item) => m4 + item * item, 0);
  const m3 = dataSet.disease.reduce((m3, item) => m3 + item, 0);
  const m5 = dataSet.disease.reduce((m5, item) => m5 + item * item, 0);

  // here u used Math.sqrt twice "Math.sqrt(m4 - m2 * m2) * Math.sqrt(m5 - m3 * m3)" 
  return (m1 - m2 * m3) / Math.sqrt((m4 - m2 * m2) * (m5 - m3 * m3));
};

console.log(corRelate());

为什么以下 Javascript 代码的输出是 NaN？

Why the output of the following Javascript code is NaN?

javascript

nan

output