Abstract
Boosting learning is a very popular learning approach in pattern recognition and machine learning. The current boosting algorithms aim to obtain a number of classifiers through learning the training samples in general. They can work well in the learning settings where the distribution of the training and testing samples is identical. The typical boosting algorithms do not pay attention to learning the test samples. They may not deal well with the large scale and high dimensional data set in which the distribution of the training and testing samples tends to be not identical. In order to deal well with the large scale and high dimensional data sets, such as face image data set, we investigate the boosting algorithms from a new perspective, and propose a novel boosting learning algorithm, termed individualized boosting learning (IBL), in this paper. The proposed IBL algorithm focuses on learning both the training and testing samples. For each test sample, IBL determines a part of training set, referred to as learning region, to perform boosting algorithm, and classify the test sample. Experiments on several popular real-world data sets show that the proposed IBL algorithm can achieve desirable recognition performance.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
More From: Optik - International Journal for Light and Electron Optics
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.