Feature selection, perceptron learning, and a usability case study for text categorization
Proceedings of the 20th annual international ACM SIGIR conference on Research and development in information retrieval
Large Margin Classification Using the Perceptron Algorithm
Machine Learning - The Eleventh Annual Conference on computational Learning Theory
Incremental Induction of Decision Trees
Machine Learning
Incremental Learning with Support Vector Machines
ICDM '01 Proceedings of the 2001 IEEE International Conference on Data Mining
An approach to incremental SVM learning algorithm
ICTAI '00 Proceedings of the 12th IEEE International Conference on Tools with Artificial Intelligence
A Second-Order Perceptron Algorithm
SIAM Journal on Computing
Online Passive-Aggressive Algorithms
The Journal of Machine Learning Research
Relaxed online SVMs for spam filtering
SIGIR '07 Proceedings of the 30th annual international ACM SIGIR conference on Research and development in information retrieval
Confidence-weighted linear classification
Proceedings of the 25th international conference on Machine learning
Lazy Bagging for Classifying Imbalanced Data
ICDM '07 Proceedings of the 2007 Seventh IEEE International Conference on Data Mining
An Online Modeling Method Based on Support Vector Machine
CSSE '08 Proceedings of the 2008 International Conference on Computer Science and Software Engineering - Volume 01
Early traffic classification using support vector machines
Proceedings of the 5th International Latin American Networking Conference
Intrusion Detection System Based on Improved SVM Incremental Learning
AICI '09 Proceedings of the 2009 International Conference on Artificial Intelligence and Computational Intelligence - Volume 01
| Hi-index | 0.00 |
In incremental learning, the classification model is incrementally updated using the small datasets. Different with existing methods, our approach updates the current classifier according to each sample in the dataset, respectively. The classifier is updated by adjusting more than the margin of each sample. Then the new classifier is generated by carefully analyzing classifier adjustments caused for labeled samples. Additionally the new classifier shall correct prediction mistakes of the previous classifier as many as possible. In details, we formulate simple constrained optimization problems and then the updated classifier is the solution derived using Lagrange multipliers. In our experiments, 13 real-world dataset are used to present the effectiveness of the proposed approach. The experimental results are shown that our update strategy is able to adjust the classifier properly. And it is also shown that the proposed incremental learning approach is suitable to be applied for the requirement of frequently adjusting the existing classifiers.