Influence functions for CART

Abstract : This paper deals with measuring the influence of observations on the results obtained with CART classification trees. To define the influence of individuals on the analysis, we use influence functions to propose some general criterions to measure the sensitivity of the CART analysis and its robustness. The proposals, based on jakknife trees, are organized around two lines: influence on predictions and influence on partitions. In addition, the analysis is extended to the pruned sequences of CART trees to produce a CART specific notion of influence. A numerical example, the well known spam dataset, is presented to illustrate the notions developed throughout the paper. A real dataset relating the administrative classification of cities surrounding Paris, France, to the characteristics of their tax revenues distribution, is finally analyzed using the new influence-based tools.
Document type :
Preprints, Working Papers, ...
Complete list of metadatas

https://hal.inria.fr/hal-00944098
Contributor : Erwan Le Pennec <>
Submitted on : Monday, February 10, 2014 - 3:43:16 PM
Last modification on : Monday, February 10, 2020 - 6:13:44 PM
Long-term archiving on: Sunday, April 9, 2017 - 10:14:20 AM

File

cart.influence.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-00944098, version 1

Collections

Citation

Avner Bar Hen, Servane Gey, Jean-Michel Poggi. Influence functions for CART. 2014. ⟨hal-00944098⟩

Share

Metrics

Record views

450

Files downloads

684