Corpus ID: 4378807

Don't Fall for Tuning Parameters: Tuning-Free Variable Selection in High Dimensions With the TREX

@inproceedings{Lederer2015DontFF,
  title={Don't Fall for Tuning Parameters: Tuning-Free Variable Selection in High Dimensions With the TREX},
  author={Johannes Lederer and C. M{\"u}ller},
  booktitle={AAAI},
  year={2015}
}
Lasso is a popular method for high-dimensional variable selection, but it hinges on a tuning parameter that is difficult to calibrate in practice. In this study, we introduce TREX, an alternative to Lasso with an inherent calibration to all aspects of the model. This adaptation to the entire model renders TREX an estimator that does not require any calibration of tuning parameters. We show that TREX can outperform cross-validated Lasso in terms of variable selection and computational efficiency… Expand
36 Citations
A Survey of Tuning Parameter Selection for High-dimensional Regression
  • 4
  • PDF
A Tuning-free Robust and Efficient Approach to High-dimensional Regression
  • 4
  • PDF
A Practical Scheme and Fast Algorithm to Tune the Lasso With Optimality Guarantees
  • 32
  • PDF
Efficient Feature Selection With Large and High-dimensional Data
  • 1
Oracle Inequalities for High-dimensional Prediction
  • 20
  • PDF
Tuning-free ridge estimators for high-dimensional generalized linear models
  • PDF
The False Positive Control Lasso
  • 1
  • PDF
...
1
2
3
4
...

References

SHOWING 1-10 OF 37 REFERENCES
Bolasso: model consistent Lasso estimation through the bootstrap
  • F. Bach
  • Computer Science, Mathematics
  • ICML '08
  • 2008
  • 355
  • PDF
Regression Shrinkage and Selection via the Lasso
  • 31,684
  • PDF
Trace Lasso: a trace norm regularization for correlated designs
  • 171
  • PDF
Stability Selection
  • 1,601
  • PDF
Uncorrelated Lasso
  • 17
Variable Selection via Nonconcave Penalized Likelihood and its Oracle Properties
  • 6,273
  • PDF
...
1
2
3
4
...