1

I need to understand the following formula:

$$u_{i}=\frac{x_i-\overline{x_i}}{{||x_i-\overline{x_i}||}_2}$$

I don't know the notation:

$${||x_i||}_2$$

Is it the second moment? Please help me to convert it to sums and number of values.

UPDATE

I need it to calculate Pearson correlation as dot product between normalised in a such a way variables. Equation 3 here.

zlon
  • 125
  • 1
  • 6
  • 4
    $|x|2$ generally refers to the [euclidean norm](https://en.wikipedia.org/wiki/Norm(mathematics)#Euclidean_norm), but as with most things it depends on context. – JMoravitz Dec 05 '17 at 15:59
  • http://mathworld.wolfram.com/VectorNorm.html – lioness99a Dec 05 '17 at 15:59
  • 2
    JMoravitz's answer is perfect. There are many ways to measure the length of a vector $x=(x_1, ..., x_n)$. The Euclidean norm defines length as $\sqrt{\sum_i x_i^2}$. Another useful definition might be $max_i |x_i|$.The correlation between two vectors has a nice interpretation as the cosine of the angle between them. That's why it's often convenient to normalize each vector to have unit length. – Eric Fisher Dec 05 '17 at 16:11
  • Thanks to every body. Sorry I have not enough reputation to vote. – zlon Dec 05 '17 at 16:22

0 Answers0