search

slds-lmu / lecture_i2ml

I2ML lecture repository
https://slds-lmu.github.io/i2ml/
Creative Commons Attribution 4.0 International
150 stars 81 forks source link

Thresholding with multiclasses #988

Open QuayAu opened 7 years ago

QuayAu commented 7 years ago

explain how with multiple thresholds you can convert a prob vector into a multiclass label

berndbischl commented 6 years ago

http://www.fharrell.com/2017/01/classification-vs-prediction.html

berndbischl commented 6 years ago

https://www.aaai.org/Papers/AAAI/2006/AAAI06-076.pdf

berndbischl commented 6 years ago

https://github.com/mlr-org/mlr/blob/master/vignettes/tutorial/cost_sensitive_classif.Rmd

berndbischl commented 6 years ago

explain the resholding properly on its own

giuseppec commented 6 years ago

Not sure if it helps but I found this paper very useful "A unified view of performance metrics: translating threshold choice into expected classification loss". However, it focusses more on threshold choice for binary (not multiclass) classification.

cem-ozturk commented 6 years ago

https://github.com/compstat-lmu/lecture_cim2/pull/583

cem-ozturk commented 5 years ago

should i take care of this issue? https://github.com/compstat-lmu/lecture_cim2/issues/585

berndbischl commented 5 years ago

Yes I assigned you but why not ask in the relevant issue thread itself 😀