Find Natural Breaks in Time Series Data

9 views (last 30 days)
I need to break my time series data at natural break points, like a Jenks Natural Break in ArcGIS. In the graph, you can clearly see 2 clusters of data in the late 1950s - 1960s and late 1970s - 1990s. I need to identify these break points statistically somehow and am not sure how.
Here is a data vector for a sample: [27 30 186 179 21 39 106 101 52 208 144 370 143 166 243 229 199 144 77 121 119 175 95 40 108 69 23 27 282 215 277 85 73 218 27 334 109 135 75 69 66 237 27 112 31 87 24 212 95 53 10 23 165 0 23 49 23 55 68 0 37 115 72 0 12 102 186 5]

Accepted Answer

QuanCCC
QuanCCC on 27 Jun 2019
It seems there's no ready-made codes for this. But you can use R (getJenksBreaks in package BAMMtools)

More Answers (1)

Guillaume
Guillaume on 28 Jun 2019
Not my domain at all, but according to wikipedia, Jenks natural break optimisation is a one dimensional k-means, which is in matlab stats toolbox, so I'd take QuanCCC statement that there's no ready-made codes for this with a grain of salt.

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!