1. What contributions have the authors mentioned in the paper "Large-scale bayesian logistic regression for text categorization" ?
The authors present a simple Bayesian logistic regression approach that uses a Laplace prior to avoid overfitting and produces sparse predictive models for text data.. The authors apply this approach to a range of document classification problems and show that it produces compact predictive models at least as effective as those produced by support vector machine classifiers or ridge logistic regression combined with feature selection.. The authors describe their model fitting algorithm, their open source implementations ( BBR and BMR ), and experimental results.
read more
2. What are the two state-of-the-art approaches used as benchmarks?
The authors also discuss two state-of-the-art text categorization approaches used as benchmarks: support vector machines (SVMs), and ridge logistic regression combined with feature selection.
read more
3. How many documents were used in the first experiment?
for efficiency the authors took a fixed, random, roughly 10% subset (77,993 documents) of the test documents as their test set in all experiments.
read more
4. How many times did the authors test the Laplace hyperparameter j?
In each run the authors tested values for the Laplace hyperparameter λj from the range 0.01–316 by multiples of √ 10, or values for the Gaussian hyperparameter τj from the range .0001–10,000 by multiples of 10.
read more
