Corpus ID: 204901837

Descriptive Dimensionality and Its Characterization of MDL-based Learning and Change Detection

@article{Yamanishi2019DescriptiveDA,
  title={Descriptive Dimensionality and Its Characterization of MDL-based Learning and Change Detection},
  author={Kenji Yamanishi},
  journal={ArXiv},
  year={2019},
  volume={abs/1910.11540}
}
  • Kenji Yamanishi
  • Published 2019
  • Mathematics, Computer Science
  • ArXiv
  • This paper introduces a new notion of dimensionality of probabilistic models from an information-theoretic view point. We call it the "descriptive dimension"(Ddim). We show that Ddim coincides with the number of independent parameters for the parametric class, and can further be extended to real-valued dimensionality when a number of models are mixed. The paper then derives the rate of convergence of the MDL (Minimum Description Length) learning algorithm which outputs a normalized maximum… CONTINUE READING

    Figures and Topics from this paper.

    Mixture Complexity and Its Application to Gradual Clustering Change Detection

    References

    Publications referenced by this paper.
    SHOWING 1-10 OF 29 REFERENCES
    Model Change Detection With the MDL Principle
    6
    Dynamic Model Selection With its Applications to Novelty Detection
    30
    Detecting gradual changes from data stream using MDL-change statistics
    11
    A Learning Criterion for Stochastic Rules
    12
    A learning criterion for stochastic rules
    78
    Decision Theoretic Generalizations of the PAC Model for Neural Net and Other Learning Applications
    852
    Minimum complexity density estimation
    494
    Information Dimension and the Probabilistic Structure of Chaos
    94
    Tracking the Best Expert
    240
    Barron and Cover's Theory in Supervised Learning and its Application to Lasso
    3