In which context is Mallow's Cp used?

Mallow's Cp is a statistical tool primarily used in the context of model selection, particularly for assessing the goodness-of-fit of regression models while considering their complexity. It helps determine the trade-off between bias and variance, ultimately aiming to select a model that balances predictive accuracy with simplicity.

By calculating Mallow's Cp, practitioners can evaluate how well a model explains the data relative to the number of predictors it includes. A lower value of Cp suggests a better fitting model, ideally close to the number of parameters plus one for an effective trade-off. This is important as using too many predictors can lead to overfitting, while too few can lead to underfitting. Thus, Mallow's Cp enables analysts to make informed decisions about which model best represents the underlying data without being excessively complex.

In contrast, the other contexts provided do not align with the specific purpose of Mallow's Cp. Feature selection in time series analysis, assessing model biases, or determining the accuracy of probability distributions involve different methods and metrics tailored to those particular tasks.