Jan Swierczek-Jereczek

Improve the performance of `PrecomputedLowfreqPowerSpectrum`

1

# Indicator summary Compute ratio between low-frequency power spectrum and total power spectrum. What is "low-frequency"? Defined by user but standard is something like 10 lowest percent of the total...

So far, `estimate_indicator_changes` and `significant_transitions` have very similar code but don't call the same, easier to maintain, functions. We should unify this such that both of them call something like:...

Handle unevenly sampled time series

5

**Is your feature request related to a problem? Please describe.** Empirical data are often unevenly sampled. In the state of the art, an interpolation is often performed but this might...

# Indicator summary Scale-averaged wavelet coeff is used to estimate the variance of high-frequency fluctuations and local Hurst exponent is interpreted as an estimator of the correlation time in the...

# Indicator summary Some pre-trained artificial neural network have shown to be able to predict transitions (and their underlying normal form). I suppose it works well if data corresponds to...

# Indicator summary Correlated noise can lead to bad transition recognition if standard CSD is applied (variance, lag1-autocorelation, AR1 parameter). Therefore use an adapted formula that addresses this. Works bad...

# Indicator summary Linearizing a 1D nonlinear model, one obtains $\dot{x} = \lambda \, x$, with $\lambda$ the restoring rate. It can be estimated in a way that is robust...