| thresh_rel | float | `0.1` | Minimum relative difference between two values to consider the latter as break candidate. See condition (1) |
| thresh_abs | float | `0.01` | Minimum relative difference between two values to consider the latter as break candidate. See condition (2) |
| first_der_factor | float | `10` | Factor of the first derivates "arithmetic middle bound". See condition (3). |
| first_der_window_size | [offset string](docs/ParameterDescriptions.md#offset-strings) | `"12h"` | Determining the size of the window, covering all the values included in the the arithmetic middle calculation of condition (3). |
| first_der_window | [offset string](docs/ParameterDescriptions.md#offset-strings) | `"12h"` | Determining the size of the window, covering all the values included in the the arithmetic middle calculation of condition (3). |
| scnd_der_ratio_margin_1 | float | `0.05` | Range of the area, covering all the values of the second derivatives quotient, that are regarded "sufficiently close to 1" for signifying a break. See condition (5). |
| scnd_der_ratio_margin_2 | float | `10.0` | Lower bound for the break succeeding second derivatives quotients. See condition (5). |
| smooth | bool | `True` | Smooth the timeseries before differenciation using the Savitsky-Golay filter |
...
...
@@ -33,7 +33,7 @@ A value $`x_k`$ of a data series $`x`$, is flagged a break, if:
* $`|\frac{x_k - x_{k-1}}{x_k}| >`$ `thresh_rel`
2. $`x_k`$ represents a sufficient absolute jump in the course of data values:
* $`|x_k - x_{k-1}| >`$ `thresh_abs`
3. Let $`X_k`$ be the set of all values that lie within a `first_der_window_size` range around $`x_k`$. Then, for its arithmetic mean $`\bar{X_k}`$, following equation has to hold:
3. Let $`X_k`$ be the set of all values that lie within a `first_der_window` range around $`x_k`$. Then, for its arithmetic mean $`\bar{X_k}`$, following equation has to hold:
