On the Interaction Effects Between Prediction and Clustering
This addresses a problem for machine learning practitioners using multi-algorithm pipelines, but it is incremental as it builds on prior work to characterize and correct specific cross-validation biases.
The paper tackles the problem of subtle adverse behaviors in cross-validation due to interaction effects between clustering and prediction algorithms, particularly in estimating out-of-cluster prediction loss with clustering errors, and provides theoretical properties and scalable estimators that correct these issues, validated on benchmark datasets.
Machine learning systems increasingly depend on pipelines of multiple algorithms to provide high quality and well structured predictions. This paper argues interaction effects between clustering and prediction (e.g. classification, regression) algorithms can cause subtle adverse behaviors during cross-validation that may not be initially apparent. In particular, we focus on the problem of estimating the out-of-cluster (OOC) prediction loss given an approximate clustering with probabilistic error rate $p_0$. Traditional cross-validation techniques exhibit significant empirical bias in this setting, and the few attempts to estimate and correct for these effects are intractable on larger datasets. Further, no previous work has been able to characterize the conditions under which these empirical effects occur, and if they do, what properties they have. We precisely answer these questions by providing theoretical properties which hold in various settings, and prove that expected out-of-cluster loss behavior rapidly decays with even minor clustering errors. Fortunately, we are able to leverage these same properties to construct hypothesis tests and scalable estimators necessary for correcting the problem. Empirical results on benchmark datasets validate our theoretical results and demonstrate how scaling techniques provide solutions to new classes of problems.