Observation-window sufficiency for churn prediction is highly design-dependent, showing a diminishing-returns knee at 45-90 days in standard setups but inverting under moving-target definitions on the KKBox dataset.
The effect of the length of the customer event history and the staying power of the predictive models in customer churn prediction: A case study of Migros Sanal Market,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
How Early Is Early Enough? Design-Dependent Observation-Window Sufficiency in Subscription Churn Prediction
Observation-window sufficiency for churn prediction is highly design-dependent, showing a diminishing-returns knee at 45-90 days in standard setups but inverting under moving-target definitions on the KKBox dataset.