Semi-supervised Learning in Support Vector Machines 1

@inproceedings{Ravi2015SemisupervisedLI,
  title={Semi-supervised Learning in Support Vector Machines 1},
  author={Sachin Ravi},
  year={2015}
}
In traditional supervised classification, classifiers are trained using feature/label pairs and the classifier performance is measured on unseen test data. In the current Internet age, as the amount of data produced grows exponentially, we would like to use as much of this data as possible to train classifiers in order to get better performance. This is especially true in models such as neural networks that have millions of parameters and thus can overfit to training sets that contain even… CONTINUE READING

Similar Papers

Loading similar papers…