pyhctsa.operations.entropy.complexity_invariant_distance¶

pyhctsa.operations.entropy.complexity_invariant_distance(y)¶

Complexity-invariant distance

Computes two estimates of the ‘complexity’ of a time series based on the stretched-out length of the lines in its line graph. These features are based on the method described by Batista et al. (2014) [1], designed for use in complexity-invariant distance calculations.

References

Parameters:¶

y : array-like¶

One-dimensional time series input.

Returns:¶

A dictionary containing the following features:

• ’CE1’float

  Root mean square of successive differences.

• ’CE2’float

  Mean length of line segments between consecutive points using Euclidean distance (Pythagorean theorem).

• ’minCE1’float

  Minimum CE1 value computed from sorted time series.

• ’minCE2’float

  Minimum CE2 value computed from sorted time series.

• ’CE1_norm’float

  Normalized CE1: CE1 / minCE1.

• ’CE2_norm’float

  Normalized CE2: CE2 / minCE2.

Return type:¶

dict