I'm writing my Bachelor Thesis within the field of neural networks and I need some preparation of the data I'm using. Do you have an idea how I am able to identify the four levels in this graph via machine learning? Especially changing points are of importance for me. In this case it would be the x-values roundabout 2700, 3700 and 4700. I already thought of Clustering, but since I'm interested of the different levels, this does not quite work out. Graph of the values to be sorted

Thankyou, Lukas

  • 1
    $\begingroup$ Why do you say the clustering would not work? $\endgroup$ – nbro Jan 20 at 22:37
  • $\begingroup$ What is the dimension of your x-axis? time? $\endgroup$ – peteR Jan 20 at 22:40
  • $\begingroup$ @nbro The cluster methods I used, did not work. For example k-means.. Do you have a cluster method in mind, that could work? $\endgroup$ – lukas_ben Jan 20 at 23:24
  • $\begingroup$ @peteR It's measured data over a period of one hour. $\endgroup$ – lukas_ben Jan 20 at 23:24
  • $\begingroup$ stats.stackexchange.com/questions/131281/… $\endgroup$ – peteR Jan 20 at 23:42

Search for "change point detection" and "change in mean" and you will find various classic textbooks and surveys on this.

Basseville, M. (1988). Detecting changes in signals and systems-A survey. Automatica, 24(3), 309-326.

Basseville, Michèle, and Igor V. Nikiforov. Detection of abrupt changes: theory and application. Vol. 104. Englewood Cliffs: Prentice Hall, 1993.

Sen, Ashish, and Muni S. Srivastava. "On tests for detecting change in mean." The Annals of statistics (1975): 98-108.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.