Corpus ID: 6967800

Interpreting Classifiers through Attribute Interactions in Datasets

@article{Henelius2017InterpretingCT,
  title={Interpreting Classifiers through Attribute Interactions in Datasets},
  author={Andreas Henelius and K. Puolam{\"a}ki and Antti Ukkonen},
  journal={ArXiv},
  year={2017},
  volume={abs/1707.07576}
}
  • Andreas Henelius, K. Puolamäki, Antti Ukkonen
  • Published 2017
  • Computer Science, Mathematics
  • ArXiv
  • In this work we present the novel ASTRID method for investigating which attribute interactions classifiers exploit when making predictions. Attribute interactions in classification tasks mean that two or more attributes together provide stronger evidence for a particular class label. Knowledge of such interactions makes models more interpretable by revealing associations between attributes. This has applications, e.g., in pharmacovigilance to identify interactions between drugs or in… CONTINUE READING

    Tables and Topics from this paper.

    Explore Further: Topics Discussed in This Paper

    A Categorisation of Post-hoc Explanations for Predictive Models
    Sparse Robust Regression for Explaining Classifiers
    Principles and Practice of Explainable Machine Learning
    A Rule Extraction Study from SVM on Sentiment Analysis
    10
    Variable importance in modern regression with application to supplier productivity analysis

    References

    Publications referenced by this paper.
    SHOWING 1-10 OF 26 REFERENCES
    Finding Statistically Significant Attribute Interactions
    4
    Analyzing Attribute Dependencies
    155
    Detecting gene-gene interactions using a permutation-based random forest method
    34
    Screening large-scale association study data: exploiting interactions using random forests
    413
    Bioinformatics challenges for genome-wide association studies
    480
    GoldenEye++: A Closer Look into the Black Box
    12