Definition:Data Coding
Definition
Data coding is a technique for simplifying manual arithmetic.
It is also useful for reducing rounding errors when using a computer.
When calculating an mean or standard deviation, the following two rules can be used:
- $(1): \quad$ If each datum is multiplied by a constant $b$, then the mean and standard deviation are each multiplied by constant $b$
- $(2): \quad$ If a constant $a$ is added to each datum, then the mean is increased by $a$ but the standard deviation is not affected.
Arbitrary Origin
Let a set of data have a data coding applied such that a constant $a$ is added to each datum.
The constant $a$ is referred to as an arbitrary origin.
Examples
Arbitrary Example
Let us wish to calculate the mean and standard deviation of the numbers:
- $179 \cdotp 385$
- $179 \cdotp 387$
- $179 \cdotp 392$
We may multiply by $1000$ and then subtract $179 \, 387$ (which is the same as adding $-179 \, 387$) to give:
- $-2$
- $0$
- $5$
The mean is:
- $\dfrac {-2 + 0 + 5} 3 = 1$
The standard deviation is:
- $\sqrt {\dfrac {\paren {-2 - 1}^2 + \paren {0 - 1}^2 + {5 - 1}^2} 3} = \sqrt {\dfrac {26} 3} = 2 \cdotp 94$
Adding $179 \, 387$ to $1$ and multiplying by $\dfrac 1 {1000}$ gives the mean as:
- $179 \cdotp 388$
Multiplying $2 \cdotp 94$ by $\dfrac 1 {1000}$ gives the standard deviation as:
- $0 \cdotp 002 \, 94$
Also known as
The technique of data coding is also referred to as just coding, but this invites ambiguity.
Hence that usage is deprecated on $\mathsf{Pr} \infty \mathsf{fWiki}$.
Also see
- Results about data coding can be found here.
Sources
- 1998: David Nelson: The Penguin Dictionary of Mathematics (2nd ed.) ... (previous) ... (next): data coding
- 2008: David Nelson: The Penguin Dictionary of Mathematics (4th ed.) ... (previous) ... (next): data coding