Distributed support vector machines

Authors:
A. Navia-Vazquez;D. Gutierrez-Gonzalez;E. Parrado-Hernandez;J. J. Navarro-Abellan
Affiliations:
Dept. of Signal Theor. & Commun., Univ. Carlos III de Madrid, Spain;-;-;-
Venue:
IEEE Transactions on Neural Networks
Year:
2006

Citing 0
Cited 8

Optimal gossip algorithm for distributed consensus SVM training in wireless sensor networks

DSP'09 Proceedings of the 16th international conference on Digital Signal Processing
Consensus-Based Distributed Support Vector Machines

The Journal of Machine Learning Research
Centered hyperspherical and hyperellipsoidal one-class support vector machines for anomaly detection in sensor networks

IEEE Transactions on Information Forensics and Security
A scalable support vector machine for distributed classification in ad hoc sensor networks

Neurocomputing
The complex fuzzy system forecasting model based on fuzzy SVM with triangular fuzzy number input and output

Expert Systems with Applications: An International Journal
The complex fuzzy system forecasting model based on triangular fuzzy robust wavelet υ-support vector machine

Expert Systems with Applications: An International Journal
Cloud-enabled privacy-preserving collaborative learning for mobile sensing

Proceedings of the 10th ACM Conference on Embedded Network Sensor Systems
Distributed machine learning in networks by consensus

Neurocomputing

Quantified Score

Hi-index	0.01

Visualization

Abstract

A truly distributed (as opposed to parallelized) support vector machine (SVM) algorithm is presented. Training data are assumed to come from the same distribution and are locally stored in a number of different locations with processing capabilities (nodes). In several examples, it has been found that a reasonably small amount of information is interchanged among nodes to obtain an SVM solution, which is better than that obtained when classifiers are trained only with the local data and comparable (although a little bit worse) to that of the centralized approach (obtained when all the training data are available at the same place). We propose and analyze two distributed schemes: a "naïve" distributed chunking approach, where raw data (support vectors) are communicated, and the more elaborated distributed semiparametric SVM, which aims at further reducing the total amount of information passed between nodes while providing a privacy-preserving mechanism for information sharing. We show the feasibility of our proposal by evaluating the performance of the algorithms in benchmarks with both synthetic and real-world datasets.