Mann Kendall Test - rileywheadon/ffa-framework GitHub Wiki

The MK Test is used to assess whether there is a statistically significant monotonic trend in a time series. The test requires that when no trend is present, the data is independent and identically distributed.

  • Null hypothesis: There is no monotonic trend.
  • Alternative hypothesis: There is a monotonic upwards or downwards trend.

Define $\text{sign} (x)$ to be $1$ if $x > 0$, $0$ if $x = 0$, and $-1$ otherwise.

The test statistic $S$ is defined as follows:

$$ S = \sum_{k-1}^{n-1} \sum_{j - k + 1}^{n} \text{sign} (y_{j} - y_{k}) $$

Next, we need to compute $\text{Var}(S)$, which depends on the number of tied groups in the data. Let $g$ be the number of tied groups and $t_{p}$ be the number of observations in the $p$-th group.

$$\text{Var}(S) = \frac{1}{18} \left[n(n-1)(2n + 1) - \sum_{p-1}^{g} t_{p}(t_{p} - 1)(2t_{p} + 5) \right]$$

Then, compute the MK test statistic, $Z_{MK}$, as follows:

$$ Z_{MK} = \begin{cases} \frac{S-1}{\sqrt{\text{Var}(S)}} &\text{if } S > 0 \\ 0 &\text{if } S = 0 \\ \frac{S+1}{\sqrt{\text{Var}(S)}} &\text{if } S < 0 \end{cases} $$

For a two-sided test, we reject the null hypothesis if $|Z_{MK}| \geq Z_{1 - (\alpha/2) }$ and conclude that there is a statistically significant monotonic trend in the data. For more information, see here.

⚠️ **GitHub.com Fallback** ⚠️