Apprentissage supervisé de règles de décision multiclasses avec contraintes de performances évolutives

@inproceedings{Jrad2009ApprentissageSD,
  title={Apprentissage supervis{\'e} de r{\`e}gles de d{\'e}cision multiclasses avec contraintes de performances {\'e}volutives},
  author={Nisrine Jrad and Edith Grall-Ma{\"e}s and Pierre Beauseroy},
  year={2009}
}
This paper proposes a method using labeled data to learn a decision rule for multiclass problems with class-selective rejection and evolutionary performance constraints. The class-conditional densities are selected from a parameterized family of estimators by maximizing the likelihood criterion. The optimal decision rule, minimizing the loss function and satisfying each of the time evolutionary constraints, is determined using the time-independent densities. A two dimensional toy problem with… CONTINUE READING

Figures and Tables from this paper.

Citations

Publications citing this paper.

References

Publications referenced by this paper.
SHOWING 1-6 OF 6 REFERENCES

Optimal Decision Rule with Class-Selective Rejection and Performance Constraints

  • IEEE Transactions on Pattern Analysis and Machine Intelligence
  • 2009
VIEW 1 EXCERPT

selective rejection rule to minimize the maximum distance between selected classes

T. Horiuchi.Class
  • Pattern Recognition,
  • 1998
VIEW 1 EXCERPT

The Optimum Class-Selective Rejection Rule

  • IEEE Trans. Pattern Anal. Mach. Intell.
  • 1997
VIEW 1 EXCERPT