`plot.oCPD()`

produces a summary plot of the results of `onlineCPD()`

. Results are somewhat hard to interpret, see Details

1 2 |

`x` |
the result of a call to |

`lines` |
if true, plot red vertical lines in the top plot to denote detected changepoints |

`title` |
the title of the plot |

`leg.name` |
the title of the legend. Note that the names in the legend are taken from the column names in |

`cleanCP` |
if true, the function will call |

`buffer` |
The minimum number of points that need to separate two changepoints before they are both replaced |

`...` |
(optional) additional arguments, ignored. |

The plotted results can be difficult to interpret.

The top plot is the data plotted as a scatterplot, with each data column in a different colour. If `lines == TRUE`

then red vertical lines will be plotted to delineate the different runs.

The bottom plot shows the algorithms results. The black smears show the probability that run length is a particular value on the y-axis at the time on the x-axis. That is, for each pair (x,y), the darkness is the probability that at time x the run length is y. If run continues, the smear continues to move diagonally up. If the run stops, the smear returns to zero on the y-axis. The red diagonal line plots the largest probability at each time. Note the log scale on Probability.

If `time`

was not `NULL`

in the call to `offlineCPD`

, then `time`

will be along the x-axis of both plots.

See `findCP`

for information on which changepoints will be removed if `cleanCP`

is true.

Depends on ggplot2, reshape2, gridExtra and scales

Zachary Zanussi

print.oCPD, summary.oCPD, str.oCPD for summary results, and `findCP`

for information on how changepoints are reduced

1 2 3 |

Questions? Problems? Suggestions? Tweet to @rdrrHQ or email at ian@mutexlabs.com.

Please suggest features or report bugs with the GitHub issue tracker.

All documentation is copyright its authors; we didn't write any of that.