Patterns

Setup

Note

  • M : number of time series

  • N : length of time series

  • R : user-defined rate of missing values (%); default = 0.2

  • W : user-defined offset window in the beginning of the series (%); default = 0.1

  • S : user-defined rate of contaminated series (%); default = 0.2


MONO-BLOCK

One missing block per series

Aligned

The missing blocks are aligned.

Note

  • R ∈ [1%, (100-W)%]

  • The size of a single missing block varies between 1% and (100 - W)% of N.

  • The starting position is the same and begins at W and progresses until the size of the missing block is reached, affecting the first series from the top up to S% of the dataset.


Disjoint

The missing blocks are disjoint.

Note

  • R ∈ [1%, (100-W)%]

  • The size of a single missing block varies between 1% and (100 - W)% of N.

  • The starting position of the first missing block begins at W.

  • Each subsequent missing block starts immediately after the previous one ends, continuing this pattern until the limit of the dataset or N is reached.


Overlap:

The missing blocks are overlapping.

Note

  • R ∈ [1%, (100-W)%]

  • The size of a single missing block varies between 1% and (100 - W)% of N.

  • The starting position of the first missing block begins at W.

  • Each subsequent missing block starts after the previous one ends, but with a shift back of X%, creating an overlap.

  • This pattern continues until the limit or N is reached.


Scattered

The missing blocks are scattered.

Note

  • R ∈ [1%, (100-W)%]

  • The size of a single missing block varies between 1% and (100 - W)% of N.

  • The starting position is randomly shifted by adding a random value to W, then progresses until the size of the missing block is reached, affecting the first series from the top up to S% of the dataset.


MULTI-BLOCK

Multiple missing blocks per series

MCAR

The blocks are missing completely at random

Note

  • R ∈ [1%, (100-W)%]

  • Data blocks of the same size are removed from arbitrary series at a random position between W and N, until the total number of missing values per series is reached.


Block Distribution

The missing blocks follow a distribution.

Note

  • R ∈ [1%, (100-W)%]

  • Data is removed following a distribution given by the user for every values of the series, affecting the first series from the top up to S% of the dataset.

To configure the distribution pattern, please refer to this tutorial.