Corpus ID: 59336361

Testing Conditional Predictive Independence in Supervised Learning Algorithms

@article{Watson2019TestingCP,
  title={Testing Conditional Predictive Independence in Supervised Learning Algorithms},
  author={David S. Watson and Marvin N. Wright},
  journal={ArXiv},
  year={2019},
  volume={abs/1901.09917}
}
  • David S. Watson, Marvin N. Wright
  • Published 2019
  • Computer Science, Mathematics
  • ArXiv
  • We propose a general test of conditional independence. The conditional predictive impact (CPI) is a provably consistent and unbiased estimator of one or several features’ association with a given outcome, conditional on a (potentially empty) reduced feature set. The measure can be calculated using any supervised learning algorithm and loss function. It relies on no parametric assumptions and applies equally well to continuous and categorical predictors and outcomes. The CPI can be efficiently… CONTINUE READING
    3 Citations

    References

    SHOWING 1-10 OF 60 REFERENCES
    A Permutation-Based Kernel Conditional Independence Test
    • 54
    • PDF
    Quantifying Uncertainty in Random Forests via Confidence Intervals and Hypothesis Tests
    • 145
    • PDF
    Kernel-based Conditional Independence Test and Application in Causal Discovery
    • 282
    • PDF
    A Scalable Conditional Independence Test for Nonlinear, Non-Gaussian Data
    • J. Ramsey
    • Computer Science, Mathematics
    • ArXiv
    • 2014
    • 26
    • PDF
    The Hardness of Conditional Independence Testing and the Generalised Covariance Measure
    • 44
    • Highly Influential
    • PDF
    Predicting good probabilities with supervised learning
    • 719
    • PDF
    PREDICTIVE LEARNING VIA RULE ENSEMBLES
    • 516
    • Highly Influential
    • PDF
    The behaviour of random forest permutation-based variable importance measures under predictor correlation
    • 188
    • PDF